THE A.I.

REVOLUTION: EVALUATING IMPACT AND

CONSEQUENCES IN COPYWRITING

Internship Report
to Universidade Católica Portuguesa to obtain a Master’s Degree
in Communication, Marketing and Advertising.

By

Veronica Provasi

Universidade Católica Portuguesa

September 2023

1
THE A.I. REVOLUTION: EVALUATING IMPACT AND
CONSEQUENCES IN COPYWRITING

INTERNSHIP REPORT

Internship Report to Universidade Católica Portuguesa to obtain a

Master’s Degree in Communication, Marketing and Advertising.

Veronica Provasi

Universidade Católica Portuguesa

Under the supervision of Alexandre Duarte

September 2023

2
Abstract
The dawn of artificial intelligence (A.I.) foreshadows the beginning of a transformative era and
prepares a profound societal shift, as witnessed in the past with historical technological innovations.
As A.I. continues to rise, its implications on content creation become more significant daily, but its
impact on copywriting remains an unexplored area in the literature.
This Internship Report delves into the understanding of the impact of A.I. on copywriting, aiming to
determine the ability of social media users to distinguish AI-generated content from human-crafted
prose.Undertaking a mixed-methods approach, this study combines the depth of qualitative insights
from focus group discussions with the breadth of quantitative data from an experimental design.
While the quantitative experiment offers an overview of participants' discernment capabilities, the
focus group discussions provide a nuanced understanding of the underlying biases, perceptions, and
emotions influencing their judgments.Results indicate a discord in discernment skill based on the
specific content pairs presented, with no overarching consensus on A.I.'s detectability, signifying
how A.I. can be recognized if used by inexperienced hands but, if used as a tool correctly, cannot be
recognized. Interestingly, participants with a writing background demonstrated a higher accuracy
rate, though other variables had minimal impact. The experiment revealed that participants
consistently underestimated A.I.'s capabilities, highlighting a gap between perceived and actual
discernment abilities.This research aimed to contribute to the nascent literature on A.I.'s role in
copywriting, emphasizing the need for a holistic understanding that integrates the technological
prowess of A.I. and the intricate human factors influencing content perception.

Keywords: Artificial Intelligence; Copywriting; Content Perception; Mixed-Methods; Focus

Groups; Experimental Design.

Resumo
O aperfeiçoamento da Inteligência Artificial (I.A) prenuncia o início de uma era transformadora e
prepara uma mudança profunda na sociedade, tal como se verificou no passado com as inovações
tecnológicas históricas. À medida que a I.A. continua a crescer, as suas implicações na criação de
conteúdos tornam-se cada vez mais significativas, mas o seu impacto na redação continua a ser uma
área inexplorada na literatura.

3
Este relatório investiga a compreensão do impacto da I.A. na redação, com o objetivo de determinar
a capacidade dos utilizadores das redes sociais para distinguir o conteúdo gerado pela I.A. da escrita
humanas. Adoptando uma abordagem de métodos mistos, este estudo combina a profundidade das
percepções qualitativas das discussões dos grupos de discussão com a amplitude dos dados
quantitativos de um projeto experimental. Enquanto a experiência quantitativa oferece uma visão
geral das capacidades de discernimento dos participantes, as discussões dos grupos de discussão
proporcionam uma compreensão matizada dos preconceitos, percepções e emoções subjacentes que
influenciam os seus julgamentos. Os resultados indicam uma discórdia na capacidade de
discernimento com base nos pares de conteúdos específicos apresentados, sem um consenso global
sobre a detetabilidade da I.A., o que significa que a I.A. pode ser reconhecida se for utilizada por
mãos inexperientes mas, se for utilizada como uma ferramenta correcta, não pode ser reconhecida.
Curiosamente, os participantes com experiência em escrita demonstraram uma taxa de precisão
mais elevada, embora outras variáveis tenham tido um impacto mínimo. A experiência revelou que
os participantes subestimaram sistematicamente as capacidades da I.A., evidenciando um fosso
entre as capacidades de discernimento percebidas e as reais.
Esta investigação pretendeu contribuir para a literatura nascente sobre o papel da I.A. no
copywriting, enfatizando a necessidade de uma compreensão holística que integre as proezas
tecnológicas da I.A. e os intrincados factores humanos que influenciam a perceção do conteúdo.

Palavras-chave: Inteligência Artificial; Copywriting; Perceção de Conteúdos; Métodos Mistos;

Grupos Focais; Design Experimental.

4
Acknowledgements:

As I approach the end of this lengthy and fulfilling academic journey, I want to thank the people
who made it possible.
First and foremost, my parents, as without their unconditional love and support, I would never have
been able to be the person I am.
The team of Sweet Legal Tech: Giulio, Carlo, and Giacomo. I couldn’t have wished for a better
team to mentor me at the beginning of my career.
Thank you to everyone who made Lisbon feel like home, in particular to Rèka and Miguel, for
always bringing a piece of my heart with you.

5
Table of Contents:
1. Introduction 12

I. Theoretical Framework 16

2. From Definitions to Deployments: An Exploration of Artificial Intelligence 16

2.1 Introduction 16

2.2 Defining Artificial Intelligence 17

2.3 Artificial Intelligence Classification: Capacities, and Potential of Different A.I. Models 18

2.4 Comparative Analysis: Artificial Intelligence and Human Intelligence 19

2.5 History of AI: The Progression and Paradigms 20

2.6 The Evolution, Rise, and Potential of Modern A.I.: Spotlight on OpenAI and GPT-3.5 22

2.7 The Pinnacle of Language Processing: GPT-3.5 Explained 24

2.8 Analyzing the Market Implications of GPT-3.5 Deployment 25

2.9 GPT-4: Unveiling the Potential and Future of Advanced Language Models 27

2.10 Conclusion 31

3. Copywriting Chronicles: The Power of Persuasive Prose 33

3.1 Introduction 33

3.2 Conceptualizing Copywriting: A Modern Perspective of an Old Art 34

3.3 Evolution of Copywriting: A Historical Journey 35

3.4 The Psychology of Copywriting 38

3.5 The Integral Role of Copywriting in Marketing and Advertising Strategies in the Social Media
Era 39

3.6 The Role of Copywriting in Shaping Brand Identity 42

3.7 Definition of Brand Personality 43

6
3.8 Copywriting as a Tool to Build Brand Personality 45

3.9 The Archetypes of Communication 46

3.10 The Utilization of the Archetypes: Case Studies 49

3.11 Conclusion 54

4. Algorithms and Creativity: How A.I. is Redefining Marketing Narratives 55

4.1 Introduction 55

4.2 A.I. and Language Models Use in Marketing and Advertising 57

4.2 Artificial Intelligence and its Effects on Consumer Behavior Patterns 62

4.3 The Revolutionary Influence of Artificial Intelligence on Written Content Creation 65

4.4 Distinction Between Human-Generated and Computer-Generated Content 68

4.5 Perception of A.I. Generated Contents 69

4.6 Ethical and Law Considerations about A.I.-generated Content 71

4.7 Conclusion 74

II. Empirical Study 76

5. Descriptive Memory of the Internship Report 76

5.1 What is legal technology? 76

5.2 Introduction to Sweet Legal Tech 77

5.2 The Structure of the Company 79

5.3 Employer Branding Strategies of Sweet Legal Tech 79

5.4 Initial Contact and Onboarding 80

5.5 My Role at Sweet Legal Tech 81

5.6 Initial Challenges and Context 82

7
5.7 Key Responsibilities During My Internship 82

5.8 Tools Utilized 85

5.9 Main Responsibilities During my Internship 86

5.10 Impact of the LinkedIn Campaigns 92

5.11 Transition and Future Endeavors 92

5.12 Critical Reflection on My Internship Experience 93

6. Methodology 95

6.1 Research questions 95

6.2 Research Design 96

6.3 Triangulation in Research 98

6.4 Target 100

6.5 Sample of Participants 101

6.6 Texts Used 103

7. Findings 109

7.1 Focus groups 109

7.2 Focus Group 1 109

7.4 Focus Group 2 113

7.5 Experiment’ Findings 117

7.6 Analysis of Correct Answers Based on Educational Degree: 120

7.8 Analysis of Correct Answers Based on Generation 121

7.9 Analysis of Correct Answers Based on Brand Content Frequency 122

7.10 Analysis of Correct Answers Based on Familiarity with A.I. 123

8
7.11 Analysis of Correct Answers Based on Confidence in Recognizing A.I. Content 125

8. Discussion: 126

8.1 Discussion of the Results of the Focus Groups: 126

8.2 General observations 126

8.2 Contrasting Perspectives 128

8.3 Differences between Generation Z and Millennials 128

8.4 Overall Analysis of the Focus Groups 129

8.5 Discussion of the Results of the Experiment: 130

8.6 Analysis of Correct Answers in Relation to Educational Highest Completed Level 130

8.7 Analysis of Correct Answers in Relation to Generation 131

8.8 Analysis of Correct Answers in Relation to Brand Content Exposure 131

8.9 Analysis of Correct Answers in Relation to A.I. Familiarity 132

8.10 Analysis of Correct Answers in Relation to Confidence and Discernment 132

8.11 Overall Results of the Experiment 132

8.12 Triangulation of Results 133

8.13 Comparative Analysis of Experimental Results and Focus Group Insights 133

9. Conclusion 136

9.1 Implications of the Research 138

9.2 Limitations 138

9.3 Suggestions for Future Research 139

10. Bibliography 140

9
List of Figures

Figure 1: The Archetypes of Communication 47

Figure 2: LinkedIn Post 1 87

Figure 3: LinkedIn Post 2 89

Figure 4: LinkedIn Post 3 89

Figure 5: Set 1, Text A Human Generated 104

Figure 6: Set 1, Text B A.I. Generated 104

Figure 7: Set 2, Text A A.I. Generated 105

Figure 8: Set 2, Text B Human Generated 105

Figure 9: Set 3, Text A Human Generated 106

Figure 10: Set 3, Text B A.I. Generated 106

Figure 11: Set 4, Text A Human Generated 107

Figure 12: Set 4, Text B A.I. Generated 107

Figure 13: Set 5, Text A A.I. Generated 108

Figure 14: Set 5, Text B Human Generated 108

Figure 15: Correct Answers to Set 1 118

Figure 16: Correct Answers to Set 2 118

Figure 17: Correct Answers to Set 3 119

Figure 18: Correct Answers to Set 4 119

Figure 19: Correct Answers to Set 5 120

Figure 20: Correct Answers x Generations 122

Figure 21: Correct Answers per Frequency 123

10
Figure 22: Correct Answers per Familiarity 124

Figure 23: Correct Answers per Confidence 125

11
1. Introduction
Throughout history, technological advancements have consistently reshaped human societies. Some
examples can be found in the invention of automobiles which, once on the market and affordable
for most of the population, entirely replaced horse-drawn carts, or the advent of electric lighting that
rendered nights much shorter and candles less essential. These innovations profoundly transformed
thousands of people's daily practices, lifestyles, and traditions, signifying broader shifts in human
thought and societal values. History is full of inventions that changed the world, fixing a present
problem and proposing a solution that appeared impossible to reach. Nevertheless, as we embrace
the new, the old does not vanish into obscurity. Instead, it becomes a testament to our past, a marker
of our evolutionary journey. This continuous interplay between the old and the new, between
memory and progress, is a fundamental aspect of human civilization, prompting us to reflect on our
origins even as we envision our future. However, it is understandable the sense of mistrust one feels
when, in everyday life, a new technology that can transform history is introduced. When OpenAI
introduced ChatGPT, an acronym for Chat Generative Pre-trained Transformer, on November 30,
2022, it was not merely unveiling another technological tool significant only for people working in
tech. This large language model-based chatbot, prepared to adapt and lead conversations according
to user preferences in terms of length, format, style, detail, and language, immediately became a
focal point of discussions far beyond the realms of artificial intelligence. Its launch sparked overall
reflection on the profound implications of A.I. for different aspects of human existence. Just as
automobiles and electric lighting once revolutionized societies, the emergence of such advanced
A.I. models immediately became one of the inventions that would change history forever and divide
the world between enthusiastic users and terrified denigrators.

In October 2022, I embarked on a six-month curricular internship journey with Sweet Legal Tech,
an Italian start-up pioneer in the field of legal technology. At the time, A.I. was something entirely
unknown to me. As I delved deeper into legal technology, I was introduced to a transformative tool:
ChatGPT by OpenAI. My superiors, recognizing its potential, suggested its application for various
minor tasks, encouraging me to use it and make my job as a copywriter and social media manager
easier. My interactions with this advanced A.I. model grew throughout my internship, allowing me
to witness its vast capacities and the revolution it indicated in the technological landscape.

12
Recognizing the profound implications of such A.I. advancements and experimenting every day
with the always more capable language model while the debate on ethics, safety, and privacy issues
became stronger and stronger, I became deeply intrigued. As a copywriter, I wondered how deeply
this powerful tool, made freely available to all, would have changed the creation of written content.
The primary research question of this study seeks to understand the future impact of A.I. on the
sphere of copywriting, an influence yet to be appropriately studied. Another question needs an
answer to understand how deeply this impact will change the copywriting field: can the average
social media user understand if a human copywriter or an Artificial Intelligence program wrote a
caption or a post? This led to my second research question and the following experiment.

This Internship Report embarks on an exploration of the profound implications of artificial

intelligence on copywriting. It comprises two main parts: Theoretical Framework and Empirical
Study. The theoretical framework presented in the literature review serves as a foundation, offering
a comprehensive background about the topics related to the research questions, as it would be
impossible to fully understand the impact of A.I. on copywriting without being aware of how LLMs
function or the importance of the tone of voice in the brand personality. The first chapter, "From
Definitions to Deployments: An Exploration of Artificial Intelligence," navigates the intricate
landscape of artificial intelligence, commencing with an explanation of the evolution of A.I.,
journeying from early tales of humanoid machines in science fiction to the groundbreaking
foundation of OpenAI and the development of GPT-3.5. The chapter further delves into OpenAI's
latest marvel, GPT-4, highlighting its potential ramifications on the market. The chapter offers the
reader a complete understanding of A.I.'s transformative potential, emphasizing its capacity to
redefine industries and societal paradigms.
The second chapter of the literature review, "Copywriting Chronicles: The Power of Persuasive
Prose", focuses on copywriting, elucidating its pivotal role in shaping brand narratives and
influencing consumer behavior. The chapter delves into the complexities of creating compelling
content, highlighting the delicate balance between creativity and strategy needed to help the brand
to reach a standpoint in the consumers' mind.

13
The third and last chapter of the theoretical framework illuminates the transformative wave of A.I.
in the creative domain, as the title "The Revolutionary Influence of Artificial Intelligence on Written
Content Creation" suggests.
The narrative underscores the potential of A.I. to generate fresh creative material, emphasizing its
growing reality in the content creation landscape. The chapter explores how A.I. is revolutionizing
content creation, from journalism to creative writing, emphasizing the implications of such
advancements in the broader content ecosystem.
Following the literature review exploration, the research provides a descriptive memory of the
internship with Sweet Legal Tech. This descriptive memory and an overview of legal technology
connect theoretical knowledge and real-world experiences. By integrating this account into the
theoretical framework, the report ensures a holistic understanding, equipping the reader with the
necessary context and background before delving into the empirical study.
This six-month internship was an incredible opportunity to apply my theoretical knowledge of
digital and social media marketing and an experience that forever transformed my vision of
technology. While exploring the diverse realm of legal technology, I gained firsthand experience
with various tools and strategies. I witnessed how they modernized the legal sector and how they
could be used in my job to facilitate the workflow, automate some smaller and longer tasks, and
optimize my time. One of these tools was indeed Large Language Models and ChatGPT.

Delving into the empirical study, the four chapters are Methodology, Presentation of results, and
Discussion, followed by the conclusion. The study employs a methodological triangulation
approach. The research design integrates qualitative and quantitative research methods, ensuring a
comprehensive understanding of the topic. The qualitative component encompasses focus group
discussions, offering a platform for an in-depth exploration of participants' perceptions regarding
AI-generated content. Simultaneously, a quantitative experiment is conducted, employing a
comparative design to measure participants' ability to distinguish between human and AI-generated
content. The research design's dual-method approach ensures data validity, reflecting participants'
accurate perceptions of AI-generated content.

14
The relevance of this research can be confirmed by the thousands of websites, applications, and
software that, in less than a year, integrated GPT-3.5 or GPT-4 in their systems, particularly tools
used for creating written content. From translations to grammar checkers, implementing A.I.
technologies is reshaping every stage of the writing process. However, whether ChatGPT could
prevail on human skills and creativity and take over the copywriter's job remains to be determined.

15
I. Theoretical Framework

2. From Definitions to Deployments: An Exploration of Artificial Intelligence

2.1 Introduction
A cloud of uncertainty has always surrounded technological development. Artificial intelligence
(A.I.), a field whose immense complexity shifts constantly and quickly, is known for these
continuous changes (Frank et al., 2019). A.I. is a highly debated subject that has emerged in every
field, from the creative halls of science-fiction novels to the meticulous analysis of academic
journals. This discussion evokes a range from wild joy to guarded skepticism. A significant point
that hangs over the discussion is the potential for A.I. to modify the structure of our labor markets
forever as we know it and establish a paradigm where computers can marginalize human functions
and cause mass unemployment (Dargham et al., 2022). Researchers have divided into two groups
in response to this issue. The first declares A.I. to be the dawn of "superintelligence" when robots
will be more intelligent than people. The second opinion, in contrast, takes a more balanced
approach, considering A.I. to be a significant but ultimately limited advancement in the history of
computer data processing (Frank et al., 2019).

This chapter aims to understand this challenging topic. The investigation brought forward in this
chapter begins by defining A.I., analyzing the various definitions offered by the previous literature
on the topic, the different kinds of A.I., and many modern applications and developments. Using
history as a guide, this research retraces the development of A.I., starting with the early stories of
humanoid machines portrayed by science-fiction and culminating with the foundation of OpenAI
and the development of GPT-3.5, a linguistic model recognized as the pinnacle of machine
intelligence in the modern era. This study also examines OpenAI's latest artificial intelligence,
GPT-4, currently the most advanced language model (GPT-4, n.d.). Despite its public release in
March 2023 (GPT-4, n.d.), it remains too early to ascertain its full ramifications on the market. As
a result, this research will focus on the influence of GPT-3.5 on the job market and the broader
economic landscape. To determine how A.I. will transform markets and redefine perceptions of
reality, the discussion is expanded to examine the significant implications of GPT-3 and analyze

16
the ethical and security concerns that emerged following the release of these powerful language
models.

2.2 Defining Artificial Intelligence

Since its birth as an academic area, Artificial Intelligence has been the center of intense debate and
discussion. It has been challenging for researchers to find a definition that captures A.I.'s expansive
and transformative character and its multifaced nature.A part of the concept of A.I. aspires to create
computers that can do jobs that have historically required human intelligence at the interface of
linguistics and computer science (Sarker, 2022). Haenlein and Kaplan (2019) describe A.I. as "a
system's ability to interpret external data correctly, learn from this data, and utilize this learning to
accomplish specific objectives and tasks through flexible adaptation" (p. 15). According to Elliot
(2022), A.I. will one day completely replicate human intellect and learning processes. Similarly,
Dobrev (2012) defines A.I. as software that, in any given circumstance, can perform comparably to
a human. On the other hand, in his book "What Is Artificial Intelligence?" published in 2007,
McCarthy defined A.I. as the science of developing clever software and intelligent devices. He
emphasizes that while it may be tempting to compare the trajectory of A.I. to that of biologically
observable human intelligence, such a framework should not limit A.I.'s use. McCarthy explains
that replicating human intellect may not be necessary, adding that problem-solving inspires A.I.
researchers regardless of biological models. Fetzer (1990) underlines the inherent difficulties in
defining A.I. in "Artificial Intelligence: Its Scope and Limits." Understanding the terms "artificial"
and "intelligence" separately may be complicated, as explained in the book chapter "What is
Artificial Intelligence?" According to Fetzer, the "artificiality" of A.I. results from human
inventiveness rather than natural processes, creating a divide between artificial and organic
intelligence. Another perspective is offered by Russell and Norvig (2010), who see A.I. as the
study of agents that react to environmental cues. These agents employ various strategies to
comprehend and respond to their environment, ranging from reactive systems to sophisticated
decision-making frameworks. The authors advocate for tailoring agent designs to specific tasks,
underscoring the significance of context. They poignantly note, "The field of artificial intelligence
attempts not just to understand but also to build intelligent entities" (Russell & Norvig, 2010, p.1).

17
The goal of A.I., according to Somalvico (1987), is not required to resemble human cognitive
abilities. In the opinion of some A.I. pioneers, this imitation is an impossible aim, especially in
light of philosophical disagreements over reality and truth. Somalvico (1987) contends that rather
than trying to reproduce every aspect of human cognition, the main goal of A.I. is to imitate certain
aspects of human intelligence to address inferential problems.

All these definitions are fundamental to understanding A.I.'s multifaceted nature and how it can be
differently applied based on its abilities. The range of abilities that A.I. systems now possess spans
from profound conceptual comprehension to learning, adaptability, and rationalization.
Additionally, A.I. can exhibit sensitivity to complex human traits, including emotions, focus,
creativity, and more (Korteling et al., 2021).

In essence, A.I. has a very complex nature, and researchers have significantly different points of
view on its goals and functions. However, it could be summarized as a series of software that aim
to create robots that mimic many aspects of human intellect, including problem-solving and
decision-making skills. Speech recognition, natural language understanding, and strategy gaming
are just a few of the activities in which A.I. systems are proficient (Holzinger et al., 2017).
Machine learning, natural language processing, and computer vision are only a few of the
technological foundations of A.I. (Elliot, 2022; Haenlein & Kaplan, 2019).

2.3 Artificial Intelligence Classification: Capacities, and Potential of Different A.I.

Models
Artificial intelligence may be categorized in specific ways based on its prowess levels and capacity
to mimic human-like functions (Elliot, 2022). According to this classification, there are four types
of A.I.: reactive machines, limited memory machines, theory of mind, and self-aware A.I. (Joshi,
2019). The most rudimentary kind of A.I. is reactive machines, which have few and no memory-
based capabilities. Automated machines are the most basic forms of A.I., with limited power and
no memory-based processing (Wang & Siau, 2019b). On the other hand, limited memory machines
can learn to make decisions based on historical data (Joshi, 2019). Cognitive A.I. is a crucial and
innovative branch of artificial intelligence designed to emulate human intelligence processes (Zhao

18
et al., 2022). Unlike traditional A.I. systems, designed for specific tasks, cognitive A.I.
encompasses a spectrum of brain-inspired cognitive functions. These include perception, learning,
judgment, reasoning, motor control, and social cognition (Zeng et al., 2023). Such capabilities
allow these systems to process and analyze data, and understand and interpret human emotions and
behaviors (Hameed et al., 2022). Cognitive A.I. research aims to develop systems that possess self-
awareness, emotions, needs, and beliefs, marking the zenith of artificial intelligence sophistication
(Joshi, 2019).

A standard distinction in A.I. is between Artificial Narrow Intelligence (ANI) and Artificial
General Intelligence (AGI) (Haenlein & Kaplan, 2019; Joshi, 2019). AGI displays the capability of
an A.I. to emulate human-like learning, understanding, and behavior, bridging multiple domains
and generalizing across them. The ultimate goal of A.I. research is to replicate facets of human
intelligence but with enhanced memory, faster data processing, and decision-making. The
progression of AGI and ANI could lead to the 'singularity,' a state that, while promising, shows
crucial potential risks to the current way of life (Haenlein & Kaplan, 2019). However, the current
stage of A.I. development remains nascent, offering ample time for research and ensuring safety
measures (Haenlein & Kaplan, 2019; Joshi, 2019).

Du and Xie (2021) underscore the transformative potential of A.I. as it becomes more ingrained in
our digital interactions. Its adaptability and learning capacity make it indispensable in the
healthcare, banking, and transportation. Despite challenges and the demand for "explainable A.I.,"
the authors remain optimistic about A.I.'s future. They highlight that foundational technologies,
such as computer vision and machine learning, drive its ongoing innovation and potential to
reshape technology."

2.4 Comparative Analysis: Artificial Intelligence and Human Intelligence

The distinction between artificial intelligence and human cognition is still the subject of heated
discussion because it has significant consequences for ethics, trustworthiness, and explainability.
An evident tendency towards anthropocentric and anthropomorphic ideals can be seen in these
discussions, elevating human-like intelligence as the goal of A.I. (Korteling et al., 2021).

19
Human thinking, an intricate tapestry constructed from interconnected brain networks, has been
primed over centuries for multifaceted perceptual-motor functions (Engelbart, 2023; Korteling et
al., 2021). Its development can be linked to a complex biological history before its appearance.
This path, primarily determined by the needs of survival and reproduction, has bred a distinctive
intellect that is especially suited yet constrained by its limitations when tackling particular fields of
non-biological knowledge (Korteling et al., 2021). In contrast, artificial intelligence, rooted in
digital computing, has unrivaled skills like data processing, memorization, and complex
mathematical computations (Korteling et al., 2021). Its increased capabilities, which now include
learning, adaptability, and even subtle responses to human traits like emotion, attention, and
creativity, indicate a limitless potential when appropriately utilized (Sallam, 2023).

A.I. is still a novice when negotiating the complex emotional tapestry, minute distinctions, and
profoundly subjective experiences that characterize human intelligence (Korteling et al., 2021).
While the potential of A.I. is generally constrained by its underlying algorithmic blueprints and
broad computing infrastructures, the inherent constraints of human cognitive skills are due to their
biological origins (Korteling et al., 2021).

2.5 History of AI: The Progression and Paradigms

The history of artificial intelligence is woven into a complex tapestry. The idea of robotic machines
acting like humans can be traced back to philosophical, literary, and fantastical roots.
Throughout history, the idea of artificial beings has captivated human imagination, leading several
famous thinkers to examine human nature using the conceptual framework of sentient machines
(Buchanan, 2005). The ancient Greeks, for instance, told tales of Talos, a bronze giant crafted by
the god Hephaestus, who acted like a human. These early stories showcased creative imagination
and touched upon the ethical challenges of creating life artificially. Adrienne Mayor's book (2018)
delves into how various ancient civilizations envisioned artificial life and human-thinking robots,
from the Greeks and Romans to the Chinese culture. These tales often reflected artisans' real-world
skills and techniques during those times. Alexandria, in particular, was a notable center for
developing advanced machines (Mayor, 2018). Transitioning to modern times, the discussion
around artificial intelligence has become more intricate. Today's philosophers and thinkers predict

20
significant societal changes driven by A.I. advancements (Jones, 2020). Ray Kurzweil, for
example, believes that the rapid progression of A.I. could drastically alter our society in the near
future. Similarly, Nick Bostrom suggests a future where a superintelligent algorithm could
dominate, highlighting its potential to tackle complex scientific and philosophical issues (Jones,
2020). In literature, Isaac Asimov's 1942 novel "Runaround" is the best example of how a novel
could conceptualize the future course of A.I. (Haenlein & Kaplan, 2019).

Nevertheless, computational and programming constraints made carrying out such projects
difficult. Nevertheless, during the past fifty years, there have been some genuinely astonishing
developments (Buchanan, 2005).Computer pioneer Alan Turing started the conversation on
artificial intelligence. He weighed how it could be possible for robots to mimic human brain
processes and consequently invented the Turing Test as a test of artificial intelligence (Russell &
Norvig, 2010). A wide range of skills, including knowledge representation, automated reasoning,
and machine learning, would be required for a computer to be successful in this examination. The
goal changed, moving from a focus on merely replicating human intellect to one on deeply
comprehending it (Russell & Norvig, 2010). Artificial neural networks, machine learning
algorithms, and natural language processing have all benefited greatly from Turing's enormous
contributions to A.I. (Buchanan, 2005).

Although Turing's contributions to A.I. are unquestionably significant, visionaries like John
McCarthy, Marvin Minsky, Allen Newell, Herbert Simon, Claude Shannon, and their
contemporaries are responsible for the field's official beginning in 1956. The theoretical foundation
for A.I. was crystallized at their symposium at Dartmouth College (Buchanan, 2005; Calo, 2017;
Pan, 2016). Periodic "summers" and "winters", times of development and setbacks, have marked
the growth of A.I. (Ilkou & Koutraki, 2020). Advances in language comprehension marked the
pioneering golden period, theorem proving, and fundamental learning (Buchanan, 2005; Brauner et
al., 2006).

The development of A.I. from the middle of the 20th century until the 1980s was supported mainly
by "symbolic A.I.". In this case, the underlying philosophy was that machines might be created

21
using rules and symbols to mimic human cognitive processes (Garnelo & Shanahan, 2019; Ilkou &
Koutraki, 2020). Tools for natural language processing and knowledge representation emerged
during this period, showing in a paradigm shift in which robots began to perform previously
thought of as exclusively human jobs. However, symbolic A.I. has limitations when faced with
ambiguous or unclear facts. This realization sparked a move towards "sub-symbolic" methods in
the following decades, ushering in the "second summer" (Ilkou & Koutraki, 2020). These methods
used neural networks and machine learning for data-driven pattern identification based on
simulating neural systems of the human brain. In order to create A.I. systems that combine human-
like reasoning with adaptive learning capabilities, symbolic and subsymbolic paradigms have come
into closer alignment in the changing environment (Garnelo & Shanahan, 2019; Ilkou & Koutraki,
2020). Recent developments in intelligent agents and autonomous vehicles highlight the need to
combine several A.I. methodologies for system effectiveness (Buchanan, 2005).

2.6 The Evolution, Rise, and Potential of Modern A.I.: Spotlight on OpenAI and
GPT-3.5
Artificial Intelligence Generated Content (AIGC), often known as material produced by A.I., has
attracted much attention lately because of the implementations of large language models (LLM)
such as ChatGPT. The OpenAI language model ChatGPT, which can communicate with users, and
the DALL-E-2 model, which creates high-quality pictures from text descriptions, are two notable
examples from major tech companies. AIGC is heading in a transformational age for A.I., with such
developments, having a significant influence. This new attention is a response to A.I.'s significant
progress, particularly with the advent of deep learning, a subfield of machine learning based on
neural networks with many layers of deep networks (Cao, 2023).

With these advancements in mind, it will be possible to delve deeper into the evolutionary journey
of A.I.. This has led to the development of a robust Artificial Intelligence system that can perform
tasks such as image recognition, natural language processing, and content generation (Dale, 2021).
However, researchers have been working on the development of A.I. for decades, and this progress
was only possible with a combination of symbolic and sub-symbolic methods that led to the third
summer of A.I. (Ilkou & Koutraki, 2020). The history of generative A.I. dates back to the 1950s, but

22
initially, progress in A.I. was more comprehensive than focused on linguistics. Growth in deep
learning coincided with improvements in generative models (Cao, 2023). Classic algorithm-based
picture-generating approaches in computer vision had their drawbacks. A turning point in the
history of A.I. was reached with the introduction of transformer architecture, which caused
convergence in generative models across several domains (Cao, 2023).
In late 2015, the course of A.I. experienced a key shift. Prominent figures in the tech industry,
including Sam Altman of Y Combinator, Greg Brockman of Stripe, Reid Hoffman of LinkedIn,
Elon Musk of Tesla, and Peter Thiel of PayPal, collectively raised one billion U.S. dollars to
establish OpenAI (Dale, 2021).OpenAI is a non-profit A.I. research and deployment organization
whos mission is to ensure that the benefits of artificial general intelligence (AGI) are accessible to
all (About, n.d.). This initiative catalyzed significant advancements in A.I. The Generative Pre-
trained Transformer (GPT) models set a new benchmark, starting with GPT-2 in February 2019.
With its 1.5B parameters and extensive training on 40 G.B. of text, GPT-2 showcased the ability to
produce coherent text resembling human speech. However, it had its limitations, occasionally
generating content that bordered on the nonsensical (Dale, 2021).
By October 2020, a collaboration between OpenAI, Stanford's Institute for Human-Centered A.I.,
and other respected institutions led to the exploration of GPT-3.5 (Tamkin et al., 2021). This model
represented a leap in A.I. capabilities, demonstrating remarkable zero-shot generalization abilities.
Despite its advancements, GPT-3.5, like all technologies, faced limitations. Concerns arose
regarding inherent biases in training datasets, leading to occasional "hallucinations" or factually
incorrect outputs that seemed plausible.
The evolution of A.I. is not limited to mere calculations. Cutting-edge chatbots are transforming
digital interactions with their ability to resonate with users emotionally (Zhou et al., 2020). In the
realm of art, algorithms like OpenAI's DALL-E and DreamStudio by Stability.ai are creating
captivating visuals, while platforms like AIVA and OpenAI's Jukebox are setting new standards in
music (Ramesh et al., 2021; Cao, 2023).
In the programming domain, tools like CodeGPT and Codex from OpenAI simplify coding tasks,
while CodeParrot assists budding programmers with constructive feedback (Cao, 2023). The
educational sector is also transforming, with platforms like Google's Minerva offering tailored
learning experiences (Loeckx, 2016; Zhai et al., 2021).

23
A.I.'s influence extends to the labor market, emphasizing the increasing value of cognitive skills and
the need for continuous skill development (World Economic Forum, 2018). As A.I. continues to
evolve, it challenges traditional notions of authorship and intellectual property, necessitating a
reevaluation of established norms in light of emerging legal and ethical dilemmas (Ballardini et al.,
2019).

2.7 The Pinnacle of Language Processing: GPT-3.5 Explained

Having explored the broader landscape of A.I., its profound impact on various domains, and the
previous versions of LLMs it is imperative to narrow our focus and delve into one of the most
groundbreaking innovations in this realm: GPT-3.5, also known as the Generative Pre-trained
Transformer 3 (Floridi & Chiriatti, 2020). GPT-3.5, developed by OpenAI, a San Francisco-based
AI powerhouse, was a new benchmark for natural language processing capabilities (Chiu, 2021;
Ray, 2020). Although it belongs to a long series of transformer types, as evidenced by its name, it
surpasses its forebears in size and sophistication. Outperforming its predecessor, GPT-3.5 is a
remarkable behemoth with 175 billion parameters and 96 layers that were painstakingly trained on
a corpus of 499 billion tokens collected from the colossal web (OpenAI.com, n.d.). In comparison,
GPT-2 seems essentially insignificant, while GPT-3.5 outperforms it by a factor of more than a
hundred (Dale, 2021). Its size is the top of the iceberg, however. The strength of GPT-3.5 comes in
its adaptability; it is capable of tasks like translation, summarization, creative writing, and even
coding (OpenAI.com, n.d.; Elkins & Chun, 2020), and it is the pinnacle of adaptable language
models. The transformer design at the center of GPT-3.5's technology allows it to sort through
voluminous data efficiently. This architecture supports its capacity to produce writing that mimics
human diction and coherence, making it a highly sought-after tool in various industries, including
banking, healthcare, and customer service (Heaven, 2021). Additionally, it is more than just a
parroting automaton; a spark of originality may be seen there, as shown by its efforts to produce
unique content (Simonite, 2020). However, like all pioneers, GPT-3.5 has its shortcomings. Its
training data, which consisted of enormous swaths of online content, occasionally led it down
biased courses, reflecting those biases. Although expressive, its outputs may need to be corrected
or appropriate. Recognizing these flaws, academics like Shin, Kim, and Cho (2021) propose
combining GPT-3.5's advantages with other machine-learning strategies to increase accuracy and

24
dependability. In particular, the possibility of bias and prejudice is raised. In addition to being
morally dubious, unmanaged discrimination in A.I. might have real-world consequences, such as
spreading prejudice or escalating social inequalities, according to Schiebinger, Turilli, and Yeo
(2021). Their request for help? A "critical intervention" in creating A.I. that ensures biases are
minimized at every stage of development. In retrospect, even if GPT-3.5 represents a significant
step forward for A.I., it is only one chapter in a longer story. The future has opportunities and
problems which require careful balancing between technical growth and moral responsibility.

2.8 Analyzing the Market Implications of GPT-3.5 Deployment

According to the McKinsey Global Institute (2018), artificial intelligence is projected to create an
economic value of $13 trillion globally by 2030, influencing diverse sectors such as healthcare,
retail, and manufacturing (Notes From the A.I. Frontier: Applications and Value of Deep Learning,
2018). In line with this, current forecasts estimate A.I.'s economic impact over the next decade to
range from $1.49 trillion to $2.95 trillion. This significant economic influence is attributed to A.I.'s
transformative applications, from interpreting biological data and automating vehicles to enhancing
fraud detection. Advancements in natural language processing have improved computer
interactions with human languages, offering potential benefits to sectors like journalism and
healthcare. Furthermore, the rise of virtual assistants and machine vision underscores A.I.'s future
omnipresence. However, the ascent of A.I. is not without challenges. While Asia and the U.S.
spearhead A.I. growth, concerns persist about the rise of dominant "super firms," increasing global
disparities, and potential wage inequalities. Effective policymaking can mitigate these concerns by
balancing A.I.'s benefits against its associated risks (Szczepanski, 2021).

The potential effects of A.I. on employment were highlighted by the World Economic Forum in
2018, highlighting the technology's capacity to automate tedious labor and increase output. The
study predicted a labor shift towards roles demanding creativity, critical thinking, and problem-
solving. By 2022, a significant portion of the global workforce will need to acquire data analysis,
social intelligence, and digital literacy skills. In contrast, manual and repetitive jobs may decline.
For instance, A.I.'s potential in the retail sector could revolutionize inventory management and
offer personalized shopping experiences. Dale (2021) elaborates on this viewpoint by imagining a

25
world rife with commercial A.I. applications in industries like renewable energy, deep space
exploration, urban planning, and disaster management. Global labor patterns will, however,
unavoidably change.

GPT-3.5, created by OpenAI, occupies a significant position within this A.I. architecture. GPT-3.5
goes beyond being a simple model because its complex design has 175 billion parameters. It
epitomizes advanced natural language processing capabilities, demonstrated by its proficiency in
text summarization, chatbot functionalities, and even intricate tasks like code generation (Tamkin
et al., 2021). Chiu (2021), Pavlik (2023), and Ray (2020) have all conducted recent studies that
support GPT-3.5's transformational impact across a variety of domains. In particular, Pavlik (2023)
highlights the potential of GPT-3.5 in journalism by showcasing features like automated news
production and content personalization. However, incorporating A.I. in journalism involves
obstacles and potential, much like most technical developments. In the past, the legal, financial,
and automotive sectors were among the first to adopt A.I., emphasizing data ownership and A.I.-
enhanced analytics as crucial differentiators (Arsenijevic & Jovic, 2019). However, the first
models' validation and improvement needed to be improved due to the lack of data in the early
phases of A.I. (Arsenijevic & Jovic, 2019).

Large Language Models, like GPT-3.5, herald transformative shifts in interaction designs.
However, they are not without limitations. For instance, their vast training data can sometimes
constrain their outputs, hindering genuine innovation (Dale 2021; Lee et al., 2023). While
GPT-3.5's outputs are predominantly accurate, they can occasionally drift into the realm of the
absurd or display biases, making the accurate representation of brand-specific nuances a challenge
(Floridi & Chiriatti, 2020). ChatGPT exemplifies these intricacies, underscoring the complexities
inherent in such models that, if applied in particular fields, can be dangerous.

In summary, A.I., especially models like GPT-3.5, drives transformative changes across industries,
influencing societal shifts, employment trends, and economic growth. The dilemma is realizing
technology's tremendous potential while intelligently addressing its complexity as it advances
inevitably.

26
2.9 GPT-4: Unveiling the Potential and Future of Advanced Language Models
In 2023, OpenAI achieved another landmark in large language models by introducing GPT-4.
GPT-4, a potent multimodal model, is skilled at handling picture and text inputs and producing text
outputs. (Carrozzo, 2023; GPT-4, n.d.) GPT-4 may still need to improve human performance in
many real-world situations, yet it excels in many professional and academic tests (GPT-4, n.d.; Katz
et al., 2023). Notably, GPT-4 demonstrated its strength by placing in the top 10% on a mock bar
exam. This is a massive improvement from GPT-3.5, which placed in the bottom 10%. (GPT-4,
n.d.)GPT-4 is, in fact, ChatGPT's most advanced language model, surpassing ChatGPT-3.5, which
was only capable of text-based communication. It makes several significant changes to the system,
as explained by Roberto Carrozzo on Il Sole 24 Ore (2023), such as the capacity to read, process,
and extract pertinent information from images, a larger capacity for information processing during
user interactions (roughly 64,000 words, as opposed to 8,000 in its predecessor), and improved
effectiveness during interactions in languages other than English. GPT-4 is an effective tool for
users and, more importantly, for its use in several commercial domains. (Carrozzo, 2023)
GPT-4 is described by OpenAi such as:

"the latest milestone in OpenAI's effort in scaling up deep learning. GPT-4 is a large multimodal
model (accepting image and text inputs, emitting text outputs) that, while less capable than humans in many
real-world scenarios, exhibits human-level performance on various professional and academic
benchmarks." (OpenAi.com, 14 March 2023)

GPT-4 was not created with a specific "assigned task" in mind, as it was common with older A.I.,
such as reading medical records or interpreting photographs. Instead, it was designed to have
general cognitive abilities to aid users in completing various activities. A prompt might be a query
or a request to do a particular action, like "Please read and summarize this medical research article."
Prompts can be written in a variety of human languages and can include data inputs like
spreadsheets, technical specifications, research papers, and mathematical formulae. They are also
not limited to being statements in the English language. (Lee et al., 2023).

27
OpenAI worked on improving GPT-4 for half a year, using knowledge from ChatGPT and
adversarial testing software. These efforts improved the model's steerability, factual correctness, and
adherence to predetermined bounds. (GPT-4, n.d.)

"We spent 6 months making GPT-4 safer and more aligned. GPT-4 is 82% less likely to respond to
requests for disallowed content and 40% more likely to produce factual responses than GPT-3.5 on our
internal evaluations." (GPT-4, n.d.)

For OpenAI, the past two years have been a period of transformation as the whole deep learning
infrastructure has undergone an upgrade. With Azure, a brand-new supercomputer was developed
that was especially suited to their workload; OpenAi defined GPT-3.5 as the inaugural trial of this
advanced system (GPT-4, n.d..) The event offered insightful information that resulted in corrections
and strengthened theoretical foundations. The GPT-4 training phase thus shows unmatched stability,
making it the first expanding model with predictable training performance.Through ChatGPT and
its API, OpenAI is prepared to make the text input features of GPT-4 available. Additionally,
OpenAi is working with a chosen partner to improve the accessibility of the image input capability.
OpenAI has also opened its OpenAI Evals framework to promote transparency and cooperative
progress. This will enable automated evaluation of the effectiveness of A.I. models.

The distinctions between GPT-3.5 and GPT-4 might appear minimal in mundane interactions.
However, as task intricacy amplifies, GPT-4's superior reliability, creativity, and nuanced instruction
handling come to the fore. To ascertain the differential capabilities of these models, they were
evaluated against various benchmarks (GPT-4, n.d.). This included emulating human-designed
exams using the latest publicly accessible tests and procuring 2022–2023 versions of practice
exams. It is pertinent to note that GPT-4 underwent no specialized training for these assessments.
While a small fraction of the exam content was exposed to the model during its training, the results
corroborated by a detailed technical report indicate its capabilities (GPT-4, n.d.).
A particular example of the capacities of GPT-4 was provided by Katz et al. (2023), who carried out
an exhaustive evaluation of GPT -4's performance on the Uniform Bar Examination (UBE), which
included multiple-choice questions from the Multistate Bar Examination (MBE) and open-ended

28
questions from the Multistate Essay Exam (MEE) and Multistate Performance Test (MPT)
components. GPT-4 performed better than its predecessors and human competitors on the MBE,
outperforming human test takers in five out of seven subjects and outperforming ChatGPT by 26%.
In addition, GPT-4 earned an average score of 4.2 out of 6.0 in the hitherto unexplored areas of
MEE and MPT, a substantial improvement above ChatGPT's outcomes. GPT-4 scored more than the
pass mark for all UBE jurisdictions when rated according to human assessment criteria across all
UBE components (Katz et al., 2023).These results, as emphasized by Katz et al. (2023), show the
exponential growth of big language models and offer a glimpse of their potential to revolutionize
industries like legal services. The UBE consists of multiple-choice and open-ended questions that
test theoretical knowledge and practical legal expertise and has always been seen as an impossible
standard for even specialized models. GPT-4's ability to uphold the standards established for human
legal practitioners across nearly all U.S. jurisdictions has disproved this prior presumption. This
demonstrates the model's exceptional aptitude for nuanced legal understanding and her reading
comprehension and writing mastery. (GPT-4, n.d.; Katz et al., 2023) Another experience of the
skills of GPT-4 was told by Katharine Sanderson (2023). As a designated red teamer for OpenAI, a
position emphasizing intensive probing for system flaws over six months, Andrew White, a
chemical engineer at the University of Rochester, has thoroughly examined GPT -4's capabilities
(Sanderson, 2023). White first noticed that GPT-4 occasionally made mistakes in atom detection
and reaction sequencing, mirroring the behavior of its predecessors while performing complex
chemical reaction prediction. However, when it was combined with external scientific resources,
such as scholarly publications, the performance of GPT-4 saw a substantial change. This integration
highlighted a key point: although large language models like GPT-4 have inherent capabilities, their
potential is greatly enhanced when linked to external digital tools like retrosynthesis planners or
sophisticated calculators, giving rise to sophisticated, novel skills. (Sanderson, 2023)

Concerns have surfaced, since the unmatched skills of GPT-4 were revealed. There is a chance that
GPT-4 will provide instructions for harmful or illegal content, especially with the knowledge of
chemical engineers (Sanderson, 2023). These reservations highlight the wider ethical ramifications
of artificial intelligence. According to Professor Luciano Floridi (2020), explicability combines the
ideas of "intelligibility" and "accountability," it is crucial for an A.I. system's ethical behavior. It

29
suggests that algorithms controlling how computers recognize patterns and make decisions should
be transparent in their design and working mechanics. With such transparency, it becomes easier to
hold systems responsible, endangering the crucial values of system fairness and privacy (Carrozzo,
2023). However, the present version of GPT-4 lacks this explicability. Users need to know how the
system functions since its technical documentation lacks transparency about architecture, dataset
origins, or operating procedures. Even though GPT-4 came from OpenAI, it is paradoxical that
nothing is known about its operating dynamics, compromising external audits of its algorithmic
function and intent (Carrozzo, 2023). There is now much more concern due to the concentration of
A.I. power in a small number of giant I.T. corporations. Despite conceding that collaborations with
these digital giants might hasten the development process, Sanderson (2023) calls for a thorough
scientific evaluation of such products. Additionally, despite its achievements, GPT-4 has several
drawbacks. The model must be made aware of current events since its expertise is limited to
historical data from 2021. This historicity might be a drawback in dynamic fields like law and
technology, where timely insights are crucial. Its comprehension could be improved by
comprehension biases, especially in specialized sectors with many unusual abbreviations.
Environmentally speaking, the fact that these models use many resources increases their carbon
footprint. In terms of safety, despite OpenAI's efforts to prevent abuse, GPT-4 is vulnerable to lousy
intent, with motivated users occasionally circumventing enforced restrictions. Pre-training data also
imprinted the model with biases, forewarning against applying it in public domains like education.
Data security and privacy have also become crucial challenges as user numbers increase. This is
highlighted by incidents like the prohibition of ChatGPT in Italy due to privacy concerns (Liu et al.,
2023).As we go farther into the era of artificial intelligence, the introduction of GPT-4 puts to the
fore a range of ethical issues that the I.T. sector and society must address. OpenAI (2023) has taken
several actions in response to the ethical and security issues around GPT-4. Integrating human
feedback, including viewpoints from ChatGPT users, is crucial in ensuring the model develops with
user perspectives in mind. Furthermore, their collaboration with more than 50 experts in A.I. safety
and security demonstrates a thorough strategy for handling possible hazards.
The safety research and monitoring systems for GPT-4 have been strategically improved by OpenAI
(2023) based on the real-world experiences of their predecessor models. Because of this dedication
to iterative improvement, GPT-4 will frequently receive changes informed by actual user

30
experiences. OpenAI has accelerated safety research by utilizing GPT-4's powerful reasoning skills,
allowing the model to produce crucial training data and improve system classifiers. In conclusion,
OpenAI (2023) demonstrates a proactive posture in bringing GPT-4 into line with the best user
engagement, safety, and security practices.

2.10 Conclusion
The preceding chapter offers a comprehensive explanation of the concept of artificial intelligence
and an exploration of the advancements in artificial intelligence, particularly emphasizing the
evolution and capabilities of models like GPT-3.5. As delineated, artificial intelligence stands at the
forefront of technological innovation, aiming to emulate human cognitive functions through
intricate computational processes or even surpass them in fields such as data processing or
memorization (Haenlein & Kaplan, 2019; Shapiro, 1992; Elliot, 2022; McCarthy, 2007; Sarker,
2022). Machines able to think and act like humans have been a recurring theme in philosophy and
literature since the ancient greek, but nowadays, such advancements are not merely theoretical;
they have tangible implications, destined to revolutionize various sectors, as emphasized by the
transformative potential of A.I. in workplaces (Zhai et al., 2021).

GPT-3.5, a product of OpenAI, emerged as a monumental stride in the realm of natural language
processing, boasting an impressive 175 billion parameters and a myriad of capabilities ranging
from translation to creative writing (Floridi & Chiriatti, 2020; OpenAI.com, n.d.; Elkins & Chun,
2020). However, like all innovations, it is full of challenges. Its vast training data, while a strength,
occasionally veers towards biases, necessitating interventions for accuracy and dependability (Shin
et al., 2021; Schiebinger et al., 2021).

The economic implications of A.I. deployment are envisaged to be enormous in the following
decades, as the McKinsey Global Institute (2018) projected, further underscoring its transformative
potential across sectors from healthcare to retail. However, A.I.'s ascent brings concerns of global
disparities, wage inequalities, and the rise of dominant "super firms" (Szczepanski, 2021).
GPT-3.5, with its 175 billion parameters, stands as a testament to the rapid advancements in AI,

31
influencing sectors from journalism to legal services (Chiu, 2021; Pavlik, 2023; Ray, 2020;
Arsenijevic & Jovic, 2019).

The unveiling of GPT-4 in 2023 marked another significant milestone, showcasing its prowess in
handling multimodal inputs and achieving human-level performance in various benchmarks
(Carrozzo, 2023; GPT-4, n.d.; Katz et al., 2023). However, with its unmatched capabilities come
concerns, especially regarding data security, privacy, and potential biases (Sanderson, 2023; Liu et
al., 2023).

The next chapter shifts from Artificial intelligence to copywriting. It underscores the integral link
between copywriting and brand identity, emphasizing the pivotal role of meticulously crafted
narratives in distinguishing a brand within the competitive realm of social media marketing
(Albrighton, 2013). The nuances of brand personality and their alignment with consumer
perceptions are further explored (Aaker, 1997; Sung & Kim, 2010). The evolution of copywriting,
especially its prominence in social media, and the strategic employment of communication
archetypes to fortify brand identity are also highlighted.

32
3. Copywriting Chronicles: The Power of Persuasive Prose

3.1 Introduction
Copywriting has been one of the most essential marketing strategies to enter customers' brains and
affect their behavior since the beginning of advertising and branding, which can be traced back to
ancient civilizations (Beard, 2017; Shaw, 2009). According to Keller (1993), brand identity is
defined by a unique combination of underlying connections, values, and symbols that collectively
establish a distinct image. A strong brand identity facilitates an emotional connection with clients,
making it more straightforward for businesses to convey their values and objectives (Kapferer,
2008). Financial success can result from a strong brand's ability to offer a long-lasting competitive
advantage (Aaker, 1996).Copywriting is essential in developing a strong brand identity by
designing messages that appeal to the audience. Businesses may build a strong brand voice in this
way, set themselves apart from rivals, and accomplish marketing goals (Albrighton, 2013).
Furthermore, copywriting may develop a strong brand personality or the set of human traits
associated with a specific brand (Aaker, 1997). As the consumer's subconscious will relate to and
forge a more vital link with a brand that may be identified as being compatible with their
personality (Sung & Kim, 2010; Mark & Pearson, 2001), it is essential to express the brand
personality in communication.

The definition, development, and significant people who have influenced copywriting from the
beginning of advertising to the social media age are examined in the following chapter. This
chapter will also examine how copywriting is used in marketing and advertising, focusing on social
media copywriting, which has grown significantly recently. Finally, it will explore how
copywriting, including various communication archetypes, may be utilized to develop a brand
identity and personality. This is essential for companies to have a distinctive and identifiable brand
image. This chapter aims to introduce copywriting, following its historical progression and
illuminating its crucial function in marketing and advertising, notably in influencing customer
behavior. The chapter also wants to critically examine the need for companies to develop a
distinctive tone of voice as a critical element of their brand identity. It explains how
communication archetypes may strengthen the brand.

33
3.2 Conceptualizing Copywriting: A Modern Perspective of an Old Art
Copywriting is a crucial element of marketing and advertising, significantly influencing the success
of brands and playing a significant role in constructing communication strategies, particularly
online (Aaker, 1996; Shaw, 2009). While there is a well-established primary definition of
copywriting in previous literature, multiple perspectives contribute to its inherent significance.
According to Albrighton (2013), copywriting involves using written content to motivate the target
audience to take specific actions. Similarly, Blynova (2018) characterizes copywriting as crafting
online content using SEO optimization, Latent Semantic Indexing copywriting, and social media
marketing to increase sales. Shaw (2009) provides a broader definition, encompassing persuasive
content creation for multiple media and settings. In summary, copywriting is a multifaceted practice
involving various forms of written communication, such as advertisements, product descriptions,
email newsletters, and any other text intended to promote or sell a product or service.

The existing literature highlights copywriting as an essential practice in marketing, attributing its
importance to its persuasive power, differentiation potential, and ability to connect with and engage
the target audience. As Shaw (2009) suggests, combining exceptional design and creative writing
can create unique communications that are extraordinary marketing tools. Effective copywriting
requires a deep understanding of the target audience and the ability to craft messages that resonate
with them, as noted by Albrighton (2013).

With the rise of digital marketing, copywriting has become more crucial than ever in enabling
businesses to stand out in crowded marketplaces. Copywriters can create compelling content that
drives sales and promotes brand recognition by establishing resonating messages and understanding
the target audience (Shaw, 2009). Businesses can create meaningful connections with their
customers by establishing a robust brand voice, differentiating themselves from competitors, and
achieving marketing objectives. These connections significantly impact consumer behavior, driving
sales and promoting brand loyalty (Aaker, 1996; Kapferer, 2008; Keller, 1993). The rise of digital
marketing ans social media marketing necessitated a reassessment of traditional marketing
strategies for businesses aiming to remain competitive in the new digital era (Baltes, 2017). This

34
shift has underscored the importance of copywriting in creating compelling brand messages that
resonate with consumers and build brand identity. Digital marketing tools have transformed the
advertising landscape, allowing consumers and producers to engage during the marketing process.
These tools have disrupted traditional marketing approaches, involving customers as active rather
than passive participants in the marketing process (Cizmeci, 2015). As the advertising landscape
evolves, businesses increasingly rely on copywriting to differentiate themselves from competitors,
establish a robust brand voice, and achieve marketing objectives (Shaw, 2009). Effective
copywriting requires a profound understanding of the target audience, the ability to construct
resonating messages, and an awareness of the various media platforms on which the messages will
be delivered (Albrighton, 2013).

By establishing a distinct and recognizable brand image, businesses can foster significant
connections with their customers, promote brand loyalty, and ultimately increase sales. Therefore,
copywriting is essential in the marketer's toolbox, enabling them to compete in a crowded
marketplace and achieve long-term success.

3.3 Evolution of Copywriting: A Historical Journey

Copywriting, the art and science of crafting persuasive text, has a long and rich history that mirrors
the evolution of advertising itself. This subchapter delves into its historical journey, highlighting
key milestones and figures.

The theoretical foundations of copywriting can be traced back to the early days of advertising when
advertisers used persuasive language to convince people to buy their products (Beard, 2017). The
first primitive forms of branding and advertising emerged in ancient societies such as Mesopotamia
and Greece, where symbols and specific adjectives were used to identify and differentiate offerings
from competitors (Beard, 2017; Maurya & Mishra, 2012) and evidence of branding and
advertisements can be found in ancient Greece and Rome, mainly related to gladiatorial exhibitions
(Sampson, 2017). Bettig (2018) traces the history of copywriting back to the Babylonian era, with
the production of the first printed material promoting a prayer book in 1477 initially, copywriting
involved creating large paper posters using ink and feathers, which were then fixed to walls and

35
poles in European cities (Maurya & Mishra, 2012). As printing technology evolved, copywriting
transitioned to smaller pamphlets and brochures. With the advent of mass printing in the 1600s,
advertisements began appearing on total pages or scattered throughout newspapers (Bettig, 2018).
In the late 1800s, John Emory Powers became recognized as the first official independent
copywriter working for Lord & Thomas and Wanamaker's department stores. He is considered the
father of modern creative advertising for his truthful and effective writing style, creating one of the
most controversial ads of the era that sold out of stock in hours (Akhter, 2018). However,
copywriting only became a recognized profession in the early 20th century (Akhter, 2018).

While the early days of copywriting laid the foundation, the 20th century, after copywriting became
a recognized profession, saw the rise of influential figures who shaped its modern form. One of the
pioneers of modern copywriting was Claude Hopkins, who wrote the book "Scientific Advertising"
in 1968. Hopkins argued that copywriting should be based on research and data rather than solely
creativity. He emphasized the significance of writing catchy headlines that draw readers into the
advertisement and understanding the consumer's needs, desires, and motivations to create a
compelling copy (Tungate, 2007).

Another influential figure in the history of copywriting was David Ogilvy, the founder of the
advertising agency Ogilvy and Mather in 1948. Ogilvy's seminal work, "Ogilvy on Advertising,"
offers a comprehensive insight into his advertising philosophy. He underscored the importance of
research, asserting that a profound understanding of the consumer is key to crafting effective
advertisements (Ogilvy, 2023). Ogilvy (2023) championed the headline's power, contending that it
often determines an ad's success or failure. He also highlighted the symbiotic relationship between
compelling visuals and copy, emphasizing that one should not overshadow the other. While he
recognized the potential of direct marketing, branding remained at the heart of his approach,
advocating for consistent messaging to build robust brand identities. His reflections made Ethical
considerations explicit, as Ogilvy believed in the advertiser's responsibility to be honest and
transparent with the consumer. This key role of honesty summurize his broader perspective on
advertising, where he emphasized the importance of understanding the psychology of persuasion to
create compelling copy. Despite being occasionally associated with the creative revolution in

36
advertising, Ogilvy was skeptical of creativity for its own sake, always prioritizing effective
persuasion (Tungate, 2007; Ogilvy, 2023).

David Abbott, the co-founder of Abbott Mead Vickers BBDO, an advertising agency established in
1977, was another key figure in the world of copywriting.Known for his comedy, simplicity, and
clarity in writing, Abbott excelled at creating catchy and powerful taglines and slogans. His impact
on the advertising industry was significant and enduring (Tungate, 2007). Emanuele Pirella is
regarded as the originator of Italian copywriting. After starting his career as a writer, he was hired as
a copywriter by chance and eventually became a junior copywriter at the American advertising firm
Young & Rubicam. Pirella collaborated with German creative director Michael Göttsche, producing
slick and humorous advertisements similar to those created by Doyle Dane Bernbach in the US
(Tungate, 2007).

In conclusion, the historical development of copywriting underlines its crucial role in the
development and effectiveness of advertising throughout the ages. From the early persuasive
language used by advertisers to the emergence of branding and advertising in ancient societies,
copywriting has consistently played a pivotal role in capturing attention, differentiating products,
and persuading consumers.

In the modern advertising landscape, where attention spans are more ephemeral, and competition is
fierce, copywriting continues to be a powerful tool. The ability to communicate persuasively, evoke
emotions, and connect with the target audience remains paramount. Copywriting has expanded its
reach in the digital age through various platforms, including websites, social media, and online
advertisements. It has become a vital component of digital marketing strategies, where a well-
crafted copy can drive engagement, conversions, and brand loyalty (Odih, 2007; Shaw, 2009). Ann
Handley (2014) elucidates the intrinsic link between proficient writing and SEO efficacy. Handley
(2014) contends that search engines favor content that melds relevance with quality. She advocates
for organic keyword integration and crafting clear, engaging narratives. Such content captivates
readers and aligns with search engine ranking criteria. The significance of evergreen content and

37
compelling meta descriptions is highlighted, underscoring their role in sustained SEO success. In
Handley's perspective, copywriting is an art and a strategic imperative in optimizing digital content.

In conclusion, it is possible to declare that copywriting remains an essential and influential aspect of
advertising. Its historical significance and continued impact highlight its ability to adapt to changing
times while retaining its essential principles. As long as there is a need to communicate and
persuade, the art and science of copywriting will continue to shape the success of advertising
campaigns, making it an indispensable discipline in marketing.

3.4 The Psychology of Copywriting

Copywriting, at its core, is the art of persuasion. It is about crafting words that inform, influence,
and motivate readers to take a specific action. The power of copywriting is in its ability to tap into
the psychological triggers that drive human behavior. Modern-day copywriting has become more
sophisticated, leveraging profound psychological principles to influence behavior (Andrews, 2023).

The principles of persuasion are essential to understand copywriting. Robert P. Cialdini (2007) a
renowned psychologist, in his book "Influence: The Psychology of Persuasio." identifies six
fundamental principles that drive human behavior: reciprocity, consistency, social proof, liking,
authority and scarcity. In copywriting, the principle of reciprocity is deeply connected to the
concept of mutual exchange, often materializing as free trials, samples, or valuable content that is
provided in return for a user's time or information (Andrews, 2023). This is closely tied to the
principle of consistency, where individuals inherently desire to remain consistent once they commit
to a particular belief or action. Modern copywriters harness this by encouraging minor
commitments, such as newsletter sign-ups, which can pave the way for more substantial future
actions (Andrews, 2023). Furthermore, the principle of social proof emphasizes the human tendency
to seek validation from others, making testimonials, reviews, and user-generated content invaluable
assets in a copywriter's toolkit (Cialdini, 2007). This is supplemented by the liking principle, which
states that generating relevant content and a resonant brand voice increases trust and likability,
making viewers more likely to respond positively. (Cialdini, 2007). Additionally, the principle of
authority highlights the importance of expert opinions, industry leader endorsements, and

38
authoritative content in bolstering a message's credibility (Andrews, 2023; Cialdini, 2007). Lastly,
the principle of scarcity capitalizes on the innate human response to urgency and exclusivity, with
tactics like limited-time offers effectively tapping into the pervasive fear of missing out (Cialdini,
2007).

Emotions also influence customer behavior, and copywriting uses this power to create captivating
narratives that resonate with the audience. Anh Hoang's study (2013) emphasizes the importance of
emotions in advertising by addressing the many aspects that evoke emotional reactions. For
example, humor can be a powerful technique for catching attention and establishing a relationship
with the audience. An advertisement's hilarious or surprising circumstance might evoke surprise,
enhancing memory and engagement (Hoang, 2013). The frequency with which we find an
advertisement on Facebook has increased significantly in the past decades to capture consumers'
attention and try to make them stand out in the multitude of advertising messages they are exposed
to daily (Davis et al., 2018). Humor in advertising draws in attention and enhances brand likability,
influencing consumer purchasing decisions. However, its influence varies depending on elements
such as previous brand ratings and the promoted goods. While comedy can help increase brand
remember, it does not necessarily immediately lead to sales. Nonetheless, its strategic use is critical
for copywriters seeking to connect with audiences and influence purchasing decisions (Lettieri,
2021).

Copywriting is a powerful marketing technique, but it demands an understanding of its psychology

to be used correctly. Persuasion principles and emotional appeal are essential in influencing
consumers' behavior and generating sales. As copywriting evolves, it is critical to remember these
psychological characteristics and apply them effectively to generate appealing and convincing
material.

3.5 The Integral Role of Copywriting in Marketing and Advertising Strategies in the
Social Media Era
Social media's rise has profoundly impacted how people interact, transforming the reality of
communication to its core, and enabling users to actively contribute to content creation rather than

39
solely retrieving information (Kissel & Büttgen, 2015). Social media encompasses various
platforms and services, including blogs, chat rooms, message boards, forums, consumer review
websites, and social networking sites (Kissel & Büttgen, 2015).

In today's advertising landscape, individuals are flooded with messages from traditional mass
media, such as television and, newspapers, and new media, such as blogs and social media (Zia et
al., 2018). In this brand saturation to which each individual is subjected daily, copywriting emerges
as a pivotal tool, crafting compelling content that stimulates consumers into action (Blynova, 2018)
and helps the consumer to remember the brand. As the advertising industry changed over time, the
essence of copywriting, freshness, innovation, and the element of surprise remains crucial for
brands to stand out (Shaw, 2012).
Historically, copywriting has been the pillar of advertising, guiding consumer choices and
encouraging brand loyalty (Zia et al., 2018). The Fordist era, characterized by efficiency-driven
administrative systems, witnessed the role of copy in molding consumer demands (Fraser, 2003, as
quoted by Odih, 2007). As market dynamics changed, characterized by an unprecedented product
variety and the growth of different niche sectors, the adaptability of copywriting became its
strength, ensuring brands connected with their intended audience (Cooper & Law, 1995, as quoted
by Odih, 2007). The 1980s, with its influx of diverse advertising channels, further underscored the
importance of impactful copy in successful campaigns (Odih, 2007).

Today's digital age presents challenges for copywriters, such as shorter attention spans and the
dominance of visual content. Copywriters must craft impactful messages, keeping into
consideration SEO optimization and real-time engagement on platforms like Twitter or Instagram
(Handley, 2014). However, its core objective remains unchanged: influencing perceptions, values,
and desires and steering consumers toward the advertised products (Shaw, 2009). Through direct
and indirect messaging, copywriting drives purchases and cements lasting brand relationships, a
testament to its enduring significance in advertising (Zia et al., 2018).

The evolving communication landscape has emphasized the significance of social media due to
globalization and the digital era's rise (Bond et al., 2010). In recent years, the importance of social

40
media has been emphasized because of the rise of globalization and the digital era, leading to
significant growth in platforms such as Instagram, TikTok, Facebook, Twitter, MySpace, LinkedIn,
and YouTube. These platforms prioritize user engagement and offer various direct communication
options, such as video, audio, postings, live chat, images, and private messaging, making them
ideal for promoting brand meaning (Kissel & Büttgen, 2015). As a result, more businesses are
advertising on these platforms, leveraging their extensive communication capabilities. However,
the uncontrollable nature of social media communication poses more challenges than the
traditional media channels (Kissel & Büttgen, 2015).

In today's competitive business world, advertising is crucial in influencing customer behavior, and
entrepreneurs rely more on technology to market their products (Applegate, 2015). Instagram, in
particular, has become a popular platform for sharing creative content with deeper meanings,
making it an essential digital asset for businesses to communicate with their target market (Jesslyn
& Agustiningsih, 2021). However, customers using social media are exposed to a multitude of
frequently contradictory information, and the presentation and content of this information can
impact their perceptions, intentions, and decisions (Sparks et al., 2013). Empirical findings
consistent with attitude formation and persuasion theories suggest that people form impressions
and beliefs based partly on the information's source. User-generated and personal information is
perceived as more informative than information derived from editorial commentary or a company's
marketing materials (Dickinger, 2011; as quoted by Sparks et al., 2013).

Recent studies highlight the growing importance of copywriting in the digital age, especially with
the intensified competition on social media (Kartsivadze, 2022). In an era where countless brands
target the same audience, the power of effective marketing is critical (Kumar & Raju, 2013). Social
media advertising seeks to shape purchasing decisions by forging memorable brand associations in
consumers' minds. These associations not only influence immediate buying choices but also foster
long-term loyalty. As consumers' emotional bonds with brands strengthen, it leads to sustained
patronage and advocacy. Thus, understanding consumer behavior and nurturing these brand-
consumer relationships are essential for enduring success (Kumar & Raju, 2013; Zia et al., 2018).

41
Raising awareness about copywriting and offering training is essential, as many still need to
integrate it into their business strategies. Mangold and Faulds (2009) suggest that while social
media influences consumer decisions, the role of copywriting in this process is equally vital.
Effective marketing strategies are crucial in today's brand-dense world. Advertising aims to create
brand memories, influencing consumer decisions. Berger and Milkman (2012) found that
emotionally evocative language in advertising can significantly sway consumer behavior. Building
and maintaining strong customer relationships is key to fostering brand loyalty (Zia et al., 2018).
"The Psychology of Advertising" by Fennis and Stroebe (2010) explores how advertising affects
consumer behavior. The authors delve into the psychological processes underlying advertising's
effectiveness in persuading people to buy products or services.

In conclusion, copywriting is essential for companies to communicate their brand's value, build
identity, and persuade consumers. It creates an emotional connection with the brand, making it
indispensable for achieving marketing and advertising objectives.

3.6 The Role of Copywriting in Shaping Brand Identity

Numerous researchers have deeply explored Brand identity in marketing (Aaker, 1996; Kapferer,
2008; Keller, 1993) as it is a fundamental tool. Keller (1993) defines brand identity as the unique
set of associations, values, and symbols that distinguish a brand, transforming it into something
special in the consumer's mind (Keller, 1993). Beyond just a name or logo, a brand encompasses
the promises a company makes to its customers and the entirety of their experience with the brand.
It's not just about fulfilling these promises but also about the holistic representation of the
company's values, purpose, positioning, and personality. A brand, therefore, is a strategic asset.
Roberts (2010) emphasizes that for a brand to be impactful, it must transcend its primary function
and symbolize something of more profound significance to its users. This distinct identity, as
highlighted by Kapferer (2008) and Keller and Lehman (2006), is a blend of tangible and
intangible elements, from the brand name and logo to its message and personality. As per Keller
(1993), a brand's name is its most vital component, encapsulating its essence and values.
Meanwhile, a logo serves as the visual embodiment of the brand, and a well-crafted one can fortify
the brand's presence in consumers' minds (Aaker, 1996).

42
Copywriting plays a significant role in shaping brand identity. A tagline, for instance, shortly
represents a brand's values and ethos. Kapferer (2008) adds that a tagline should be catchy,
memorable, and encapsulate the brand's positioning. Aaker (1996) emphasizes that the language
used to describe the brand and its messaging should be consistent and mirror the brand's character.
Consistency in brand representation across all marketing channels is crucial, reflecting its core
values and mission (Kapferer, 2008). By utilizing these elements, businesses can establish visual
and emotional identities, resonating with their audience, enhancing brand equity, loyalty, and,
ultimately, financial performance (Keller, 1993). In today's competitive landscape, a robust brand
identity, shaped effectively through copywriting, is indispensable for sustainable growth and
profitability (Aaker, 1996).

3.7 Definition of Brand Personality

Researchers have long proposed that most brands possess personality traits (Aaker, 1997) and
symbolic qualities (Kissel & Büttgen, 2015). Brand identity is intimately linked to brand
personality (Kissel & Büttgen, 2015), which refers to the collection of human characteristics
closely associated with a particular brand (Aaker, 1997, p. 347). Unlike product-related attributes
that primarily fulfill utilitarian purposes for consumers, brand personality tends to fulfill a
symbolic or self-expressive role (Keller, 1993). This means that customers often ascribe human-
like personality traits to companies, enabling the symbolic use of brands (Aaker, 1997).
Keller (1993) argues that consumers often attribute human-like personality attributes to companies.
Thus, consumers perceive brands as entities with distinct personalities capable of evoking
emotions and establishing connections. As a result, brands become symbolic representations that
individuals can use to express themselves or communicate specific values or aspirations. A brand
with a strong personality that embodies positive attributes such as honesty, reliability, and quality
can resonate with customers on a deeper level (Keller, 1993). Similarly, job seekers also attribute
personality characteristics to potential employers. These characteristics are often considered
subjective, abstract, and intangible attributes of the organization. The perception of these attributes
can play a significant role in the job seeker's decision-making process (Kissel & Büttgen, 2015).
Therefore, companies must develop a strong brand personality that aligns with their target
audience's values and aspirations. This is a clear sign of how brand personality is a vital component

43
of brand identity, as perceived by the target audience within the brand image. By establishing a
brand personality that resonates with the target audience, companies can create a stronger
connection with their customers and employees, increasing brand loyalty and overall brand
reputation.
Brand personality is crucial for marketers and consumers at the same time. It forms an essential
part of a brand's image and value in consumers' minds, contributing to its equity (Keller, 1993). A
distinctive, strong, desirable, and consistent brand personality can foster a connection with
consumers, enabling marketers to differentiate their brand from competitors and enhance
marketing effectiveness. From the consumer perspective, brand personality offers self-expressive
and symbolic benefits (Aaker, 1999). Consumers often use brands as a means of expressing and
communicating their personalities. By linking their self-concepts with brand images and forming
self-brand connections, consumers satisfy their needs and establish brand equity, fostering long-
term relationships (Sung & Kim, 2010). Additionally, associating with brands provides
psychological benefits by helping consumers shape and define their self-concepts and present
themselves in various social contexts (Escalas & Bettman, 2003, as quoted by Sung & Kim, 2010).
According to Sung and Tinkham (2005), brand and human personalities are only partially
comparable. In psychology, personality refers to consistent behavioral tendencies in various
situations known as "traits." These traits are descriptive labels psychologists and sometimes
laypersons use to characterize an individual's behavior, physical attributes, attitudes, beliefs, and
demographic characteristics. In contrast, brand personality is not aligned with the concept of
human personality. Brands do not engage in actions, and their personality is shaped by physical
attributes, functional performance, characteristics of brand users, and the social or physical
contexts in which the brand exists. Additionally, brands may hold symbolic meaning, such as
prestige or group identification. Sung and Tinkham (2005) argue that while human personality
traits comprise both implicit (perceived) and objective (independent of perception) components,
brands lack objective personality traits separate from a consumer's perception.
Keller and Richey (2006) proposed a comprehensive model of corporate brand personality that
includes three key elements: affective (emotions), cognitive (thoughts), and conative (behaviors)
dimensions. These elements are divided into three categories: heart, mind, and body. Passion and
compassion are the company's core values, while the body comprises agility and teamwork.

44
Furthermore, the company's intellect is characterized by innovation and discipline. A strong brand
personality is crucial for any business as it can foster enthusiasm, empathy, creativity, discipline,
agility, and teamwork. By emphasizing these personality attributes, companies can differentiate
themselves from their competitors and build strong customer relationships. In addition, a
company's brand personality can influence the behavior of its employees and inspire them to work
towards a common goal. (Keller and Richey, 2006). In conclusion, a business that invests in
developing a strong brand personality can reap numerous benefits. By aligning the company's core
values, behaviors, and thoughts, an organization can create a cohesive and compelling brand
personality that resonates with its target audience.

3.8 Copywriting as a Tool to Build Brand Personality

Copywriting plays a pivotal role in shaping a brand's personality, helping the brand resonate with
its target audience. Beyond tangible elements like logos, the subtle language of copywriting
encapsulates a brand's core values and strengthens the emotional connection created with its
customers (Aaker, 1997).

The relationship between brand personality and copywriting is deep, as copywriting serves as a
conduit for brands to articulate their distinct personality, values, and beliefs (Bettig, 2018). Tungate
(2007) underscores the power of copywriting in weaving a compelling brand narrative that
captivates consumers and fosters a genuine connection. This connection, rooted in language that
aligns with the brand's values, is instrumental in creating a unique brand identity in a competitive
landscape (Kapferer, 2008). Holt (2004) further elaborates on the significance of this alignment,
emphasizing that a brand's values guide its actions and decisions. By resonating with these values,
copywriting can amplify the emotional bond with consumers, a cornerstone for cultivating brand
loyalty. Kapferer (2012) asserts that brand identity is an amalgamation of core values, external
expressions, and strategic marketing endeavors. Successful brands are adept at crystallizing a
compelling identity and seamlessly conveying it to their audience. Within this framework,
copywriting emerges as a vital tool for external expression, shaping perceptions and solidifying
brand identity. A consistent tone of voice, as highlighted by Aaker (1997), further cements this
identity, fostering a sense of brand familiarity and loyalty.

45
Venturing into psychology, Sung and Tinkham (2005) draw a nuanced distinction between human
and brand personalities. While human personalities are an amalgamation of consistent behavioral
traits observable across a spectrum of situations, brand personalities are sculpted by many factors,
including attributes, performance metrics, user demographics, and prevailing contexts. Intriguingly,
brands, unlike humans, derive their personality exclusively from the lens of consumer perceptions.
Often laden with symbolic meanings such as prestige, innovation, or trustworthiness, these
perceptions play a pivotal role in shaping brand-consumer interactions and relationships.

In conclusion, brand personality, anchored in relatability, significantly influences consumer

behavior. Marketers harness diverse strategies, from narratives to emotional appeals, to forge a
robust connection with their audience. This connection, when nurtured, translates into brand
loyalty, increased sales, and sustained success.

3.9 The Archetypes of Communication

Much thought has been given in marketing and consumer psychology to the idea that inanimate
objects, like commercial brands, might associate themselves with human traits. However, as the
brand's goal is to enter consumers' minds, brand personalities must use more profound symbolic
meaning (Sung & Kim, 2010). To do this, they rely on archetypes.

Carl Jung, a notable Swiss psychiatrist and psychoanalyst, originated the concept of psychological
archetypes (Siraj & Kumari, 2011). Jung, renowned as a pupil of Sigmund Freud and a key figure
in the development of psychoanalysis, conducted extensive research encompassing various
disciplines such as anthropology, literature, philosophy, psychology, and religious studies (Finzi,
2017). While exploring cultural myths, he observed that all legends exhibited specific, universally
recognized characteristics (Siraj & Kumari, 2011). Archetypes represent personifications of
behaviors— characters who embody behavior patterns (Jung,1964, 1968, as quoted by Faber &
Mayer, 2009). While symbols can manifest in various forms, Jung's theory of archetypes finds its
inception in recognizing certain symbols possessing an objective value (Finzi, 2017).

46
 Figure 1: The Archetypes of Communication

As defined by Silvia Finzi in her book "Storia della psicoanalisi" [History of psychoanalysis]
(2017, p.314):
"The archetypes are original images that participate in the instinct, feeling, and thought, while
retaining their autonomy; they are the memory of humanity that remains in the unconscious. However, it is

47
a collective unconscious, a matrix common to all peoples, regardless of time and place, a virtual image of
the world transmitted by genetic inheritance.”

According to Jung, who defined the concept, archetypes underlie the most influential hypotheses
throughout history (as quoted by Siraj & Kumari, 2011.) Archetypes are fundamental since they are
universal and intuitive and can be found in myths, stories, symbolism, rituals, and human impulses.
As a result, archetypes can offer insightful knowledge regarding how people react to other people,
fictional characters, and media. (Faber & Mayer, 2009). Applying archetypes as universal symbols
has relevance in the communication domain, particularly in developing a brand personality. The
psychology of archetypes, which includes images, symbols, characters, and stories that are always
present and convey messages in our collective unconscious, significantly impacts how a company
develops its visual identity and positioning (Baronio et al., 2022). Bechter et al. (2016) established
a connection between Jung's archetypes and brand personality, highlighting the significance of
aligning communication strategies and brand personality traits to resonate with clients. Just as
individuals select brands that reflect their communication style and personality, consumers choose
brands that align with their perception or desired image of a product. (Sung & Kim, 2010)
Faber and Mayer (2009) expanded upon Jung's theory by developing the Neo-archetypal theory,
which outlines five critical characteristics exhibited by archetypes:
1. Archetypes manifest as story characters while being represented psychologically as mental
models such as self and other schemas and prototypes.
2. Encountering archetypes often elicit intense emotional responses.
3. Archetypes operate at an automatic or unconscious level.
4. Archetypes are culturally enduring, facilitating easy learning and widespread recognition.
5. Archetypes serve as "mental standard models" that can evoke strong emotional reactions in
individuals exposed to them.
The automatic operation of particular mental models is another noteworthy feature of the Neo-
archetypal hypothesis. These models are usually processed implicitly or automatically, affecting
how people see themselves and others and interpret their feelings. Researchers Faber and Mayer
(2009) found that people react differently to different archetypes depending on their personalities
in their study on resonance to archetypes in media. Understanding archetypes can provide practical

48
insights into an individual's personality, preferences, and dislikes. Mark and Pearson (2001)
emphasize that customers are drawn to brands that align with major or emerging archetypes in their
subconscious. Therefore, comprehending an individual's "archetypal life themes" could prove
crucial in targeted communication strategies, especially as personalization becomes increasingly
significant in advertising. Further, archetypes play a pivotal role in storytelling, a fundamental
aspect of human nature. Stories possess immense power in shaping our worldview, and the
presence of archetypes in storytelling helps construct fictional narratives (Ganassali &
Matysiewicz, 2021; Severin & Tankard, 2001).

Brand storytelling is a dynamic branding strategy that involves presenting ongoing stories about
the business by disseminating helpful brand-related content to customers. Publishing material now
encompasses more than just traditional print media and incorporates electronic and social media
platforms in various media formats. By generating customer experiences through material relevant
to their brands, marketers interact with consumers in this way. (Du plessis, 2015)

3.10 The Utilization of the Archetypes: Case Studies

Storytelling is a vital human activity because it serves as the primary means through which humans
create and comprehend world representations and then communicate those views. Archetypes help
understand the intrinsic meanings of brands and products, thereby assisting marketers in
establishing lasting concepts and identities that resonate with consumers' most basic emotions and
motivations. This connection is crucial for creating loyal consumers and motivating sales (Baronio
et al., 2022).

Each archetype represents a particular worldview according to various life objectives and views on
what gives life purpose (Pearson, 1989). Archetypes, then, offer a common language that describes
what people strive to experience and repeat, whether consciously or unconsciously (Woodside et
al., 2008). Since they describe the models already present in human brains and offer authentic,
emotionally realistic stories, even when portrayed through incredible or impossible events, myths,
and fairy tales serve as vehicles of coded meaning and representations of these archetypes
(Zehnder & Calvert, 2004). Brands have been urged to employ brand archetypes to create effective

49
communication strategies and solidify consumer relationships (Ganassali & Matysiewicz, 2021)
but to understand the utilization of archetypes in branding and advertising, we must educate
ourselves with the more common archetypes and examine examples of brands that have effectively
employed them. Luis Raul Jasso offers an overview of the twelve most popular archetypes used in
branding, shedding light on their characteristics and significance (Jasso, 2013).

One such archetype is "the Explorer." Motivated by a strong desire for new adventures, Explorers
are willing to go to great lengths to escape monotony and confinement, embracing the unknown.
They embody a spirit of adventure and are driven to seek new insights and expand their horizons.
Jeep is an excellent example of a brand that effectively embodies the Explorer archetype. Through
its advertising campaigns, Jeep communicates a sense of adventure, encouraging viewers to
explore the unknown while enjoying their Jeep vehicles' comfort and capabilities (Jasso, 2013).
Another example is the use of the archetype "the Magician," perfectly explained by Disney. For
instance, Disney promotes itself as mystical, enigmatic, and alluring when it employs the tagline
"Where Dreams Come True." The tagline and associated marketing materials offer enough of a hint
for customers' subconscious minds to fill in the blanks and draw from a mental archive of
archetypes to build the rest of Disney's narrative (Merlo et al. l, 2022). Another contrasting
archetype is "the Lover." The Lover archetype, rooted in love, physical pleasure, and sensual
experiences, expresses unwavering appreciation and passion for beauty, intimacy, and partnership.
Their perspectives are often influenced by emotions that transcend time and manifest through
sensations, culminating in an inexplicable belief that love is the world's most fulfilling experience
(Jasso, 2013).

Another study is "Effects of Brand Personality on Brand Trust and Brand Affect" by Sung and Kim
(2010). This study proposes that specific dimensions of brand personality have a stronger
relationship with brand trust, while others are more closely associated with the brand effect. The
findings of this research indicate that the Sincerity and Ruggedness dimensions of brand
personality are more influential in shaping brand trust than brand effect. Conversely, the
Excitement and Sophistication dimensions have a stronger connection with brand effect than brand
trust. The Competence dimension exhibits similar effects on both brand trust and brand effect.

50
These results align with the assertions of marketing and consumer researchers, suggesting that
brand personality can enhance brand trust, evoke brand effect, and ultimately foster brand loyalty.
By understanding and tapping into these archetypes, brands can effectively communicate with their
target audience, eliciting specific emotions and associations that resonate with consumers' desires
and aspirations because the typical characteristics of characters and plot lines in myths, fairy tales,
books, and films are examples of archetypes, universal themes of human existence (Faber &
Mayer, 2009).

While archetypal figures do not strictly adhere to gender categorizations, specific characteristics
and traits often emphasize associations with one gender over another (Roberts, 2010). For instance,
archetypes such as The Lover, The Caregiver, and The Innocent are traditionally associated with a
female audience. These archetypes embody compassion, stability, and closeness, commonly
perceived as more aligned with feminine attributes (Faber & Mayer, 2009, p. 308).
Conversely, archetypes such as The King, The Warrior, The Magician, and The Hero have
traditionally been associated with the male psyche (Moore & Gillette, 1991). These archetypes
exemplify characteristics typically associated with adult males, reflecting strength, leadership,
power, and bravery. Mark and Pearson (2001) demonstrated how advertisements for cigarettes and
autos frequently employ the Hero archetype. Roberts (2010) conducted research revealing various
dominant archetypes associated with different product categories. For instance, sports drinks and
sports were associated with the "hero" archetype, automobiles and athletic shoes with the
"explorer" archetype, video game consoles with the "jester" archetype, beauty products with the
"lover" archetype, soft drinks and beer with the "everyman" archetype, insurance with the
"caregiver" archetype, energy drinks with the "outlaw" archetype, apparel with the "ruler"
archetype, and political parties with the "ruler" archetype.

Another fascinating example of how an archetype attracts a specific kind of brand or a specific
product is the case of beauty products targeted toward a feminine audience, often employing
archetypes such as The Caregiver or The Innocent. This is to convey traditional female values and
ideals. Drawing from psychological theories on archetypes, this phenomenon can be understood as
a strategic marketing approach to resonate with the desired consumer base (Moore & Gillette,

51
1991).In addition, Jasso (2013) explains how using archetypes such as The Caregiver or The
Innocent for beauty products reflects an understanding of societal expectations and ideals
traditionally associated with femininity. The Caregiver archetype, characterized by nurturing,
empathy, and compassion, aligns with the nurturing and self-care aspects often emphasized in
beauty product advertising. Similarly, The Innocent archetype represents purity, youthfulness, and
vulnerability, which can evoke a sense of beauty, innocence, and the desire to maintain a youthful
appearance. Faber and Mayer (2009) support this notion in their study "Resonance to Archetypes in
Media: There's Some Accounting for Taste," exploring the resonance of archetypes in media and
their influence on individual preferences. They argue that individuals naturally resonate with
specific archetypes based on their personality traits and preferences. Beauty product advertisers
leverage this resonance by strategically choosing archetypes that align with their target market's
desired values and aspirations, thus enhancing the appeal and connection to their products.
Beauty product advertising uses archetypes like The Caregiver or The Innocent to connect with
female consumers emotionally. Advertisers tap into collective patterns and cultural narratives
associated with femininity to evoke emotions, reinforce values, and align products with the desired
consumer identity.

Although archetypes are associated with certain genders, individuals of all genders can resonate
with and embody various archetypal traits. Recognizing gender archetypes enables a nuanced
analysis and effective communication strategies that appeal to specific audiences (Moore &
Gillette, 1991; Roberts, 2010). The utilization of archetypes is a fascinating phenomenon that
permeates our daily lives, often operating on a subconscious level. Despite their subtle presence,
archetypes' influence on consumer choices is undeniable. Researchers have extensively analyzed
how brands can possess personalities akin to individuals (Azoulay & Kapferer, 2003). This concept
is of great significance to both businesses and consumers.

Another research that delve deeply into the analysis of the impact of archetypes was conducted by
Baronio et al. (2022), who evaluated the effectiveness of using archetypes in developing logos for
a design agency. The researchers chose three archetypes for the study: the Mage, who symbolizes
transformation and change; the Jester, who represents pleasure and present-moment living; and the

52
Creator, who represents invention and creation. These archetypes served as the basis for creating
three distinct logos, each of which was created to communicate the idea and traits associated with
its specific archetype. The next step was to poll the brand's target market to ascertain whether the
logos effectively communicated the desired archetype's message and established a positioning for
the agency. The study's findings showed that it is feasible to create a visual identity based on
archetypal ideas that effectively evoke joyful feelings and sensations in the target market (Baronio
et al., 2022). This research implies that a brand's personality may be significantly shaped by the
deliberate application of archetypes in copywriting and visual identity creation. A brand may
develop a more engaging, believable, and noticeable identity by matching its visual components
with archetypes that speak to the emotions and motives of its target audience. As a result, there
may be deeper bonds with the audience, more customer loyalty, and a more clear positioning in the
market.

Brand personality is crucial in transforming a mere object into an iconic symbol. By imbuing
brands with distinct personalities, their meaning transcends the utilitarian nature of mere tools or
products (Aaker, 1997), and they can gain a self-expressive role (Keller, 1993). It is through
establishing a brand personality that a more profound connection is formed with consumers,
evoking emotions, building trust, and fostering brand loyalty. (Kissel & Büttgen, 2015)
Recognizing the power of archetypes and harnessing them allows businesses to align their brand
personalities with the desires and aspirations of their target audience. This strategic alignment is
essential for establishing a solid and resonant brand identity in the minds of consumers (Sung &
Kim, 2010). When consumers can identify the origins of a product, they often attribute
responsibility to specific manufacturers or distributors, thus holding them accountable for the
product (Keller, 1998, as quoted by Roberts, 2010). This attribution of responsibility, coupled with
the communication efforts undertaken by the brand, can cultivate a sense of belonging among
consumers, forging a connection between them and the brand itself. (Roberts, 2010; Sung & Kim,
2010) By aligning with specific archetypes, brands can establish a cohesive and enduring identity
that resonates with the collective consciousness of their target audience (Maurya & Mishra, 2012).
Brands may use archetypes to develop a unified story that evokes feelings, grabs attention, and
establishes a unique position in the market.

53
In conclusion, employing archetypes enables businesses to sustain a significant and meaningful
presence in customers' thoughts. Using archetypes, brands may successfully encourage consumer
familiarity, trust, and loyalty. They may then ingrain themselves into the cultural fabric, influencing
customer perceptions and preferences, and penetrating the communal consciousness.

3.11 Conclusion
This chapter emphasized the value of copywriting persuasive language in advertising and
marketing and how it is crucial for brands today to stand out from their rivals (Applegate, 2015;
Shaw, 2009; Zia et al., 2018), especially on social media (Kissel & Büttgen, 2015). The idea of
brand personality is crucial for turning everyday items into iconic symbols because it gives
businesses distinctive qualities that go beyond their purely functional nature. (Aaker, 1996, 1997;
Keller, 1993; Kapferer, 2008). Businesses may develop a stronger connection with customers,
inspire feelings, generate trust, and encourage brand loyalty by developing a brand personality
(Keller, 1993; Kissel & Büttgen, 2015). Businesses can create a solid and resonant brand identity
that resonates with customers' collective awareness and preferences thanks to this alignment of
brand personality with the needs and goals of the target audience (Sung & Kim, 2010).
Additionally, companies may connect their brand identities with the needs and goals of their target
audience by understanding the power of archetypes and using them (Kissel & Büttgen, 2015;
Roberts, 2010). This encourages familiarity, trust, and loyalty. Advertising and brand
communication that strategically employs archetypes may help businesses infiltrate the collective
unconscious, embed themselves in the cultural fabric, and influence customer perceptions and
preferences (Maurya & Mishra, 2012). Ultimately, the literature examined in this chapter
demonstrated that utilizing compelling language and archetypes enables businesses to sustain a
significant presence in customers' thoughts over time, resulting in enhanced brand equity, brand
loyalty, and improved financial performance.

54
4. Algorithms and Creativity: How A.I. is Redefining Marketing Narratives

4.1 Introduction
Especially in light of language's origins and nature, the distinction between humans and animals
has long been the topic of considerable discussion. Divergent viewpoints emerged from various
academic and ideological backgrounds; historians, linguists, anthropologists, and protectionists all
contributed their distinct points of view (Pearson, 2013). Was language merely a biological
function, an improved form of the gestures used for communication by animals? Or was it a unique
manifestation of the human spirit that set us apart from the rest of the animal kingdom? The origin
of language also came into question: Was it a construct of deliberate design or the result of slow
evolution? Ultimately, the question at the center of this decades-long debate is whether speaking is
the primary characteristic that makes us human. Does the unmistakable sign of human intelligence
lie in the ability to create logical stories and creative tales through the written word? (James, 1878;
Pearson, 2013)

Nevertheless, artificial intelligence (A.I.) systems nowadays can create complex creative and
literary works without human authorial input (Racine, 2018). Previously, natural language
generation was based on rigid grammar systems, basic statistical models, and extensive guidelines.
Their capacity was relatively constrained; it frequently resulted in jumbled phrases or
conversations restricted to a few topics. Large-scale statistical models, like GPT-3.5, have
revolutionized this discipline. GPT-3.5, in contrast to conventional models, learns the subtleties of
language not through rigid rules but through extensive experience (Elkins & Chun, 2020). A.I.
software such as GPT-3.5 and GPT-4 are not created with a specific assigned task in mind.
However, it was designed to have general cognitive abilities to aid users in completing various
activities. (Lee et al., 2023) This allowed them to perform tasks that, prior to the inventions of this
kind of A.I., were only human prerogatives, such as creating original creative content. Artificial
intelligence advancements have made computer-generated material a reality (Dargham et al.,
2022).
The rapid development and proliferation of the internet and new technologies like artificial
intelligence have significantly reshaped the marketing and advertising industry.

55
These changes have necessitated that brands evolve their methods of interaction with customers.
The role of copywriting, the art of crafting persuasive and engaging language to promote a product
or service, has been highlighted in this transition. Duarte and Neumaier (2022) underline the
importance of a copywriter's ability to create content that captivates attention while being
informative, persuasive, and straightforward. However, with A.I. now playing a growing role in
copywriting, questions about the future of human creativity in an automated world have surfaced.
The worry that automation and robots could replace human jobs is prevalent, though not
necessarily justified (Dargham et al., 2022).

"When cars began to take over from horses, horse feeders didn't have to starve. They mostly got
other jobs. The rise of robot writers won't necessarily make human writers obsolete. If humans don't write
for free, the number of words published will decline by some percentage. If the percentage is large enough,
the decline in the number of written words will be larger than the decline in quality, and total information
output will decline." (Dargham et al., 2022, p.1)

As witnessed in the past, workers often adapt to other professions rather than becoming jobless.
The advent of A.I. writers does not imply the end of human writing. If payment for human authors
stops, there may be a decrease in the amount of written material that is generated, which might be
more than any potential fall in quality, hence lowering the amount of information produced overall
(Dargham et al., 2022). Businesses must adjust proactively to remain competitive as this market
evolves (Duarte & Neumaier, 2022). Copywriting, an essential aspect of marketing and advertising
that uses persuasive and engaging wording to support a product or service and entice potential
buyers (Aaker, 1996), must also evolve and adjust. Effective copywriting involves a thorough
grasp of the target audience as well as the aims of the organization (Albrighton, 2013). It is
essential in various marketing communication platforms, including websites and social media,
product descriptions, email marketing campaigns, and commercials (Shaw, 2009). Copywriting
aims to raise brand recognition, increase brand awareness, and convert potential buyers into loyal
customers. An already strong brand identity can still be strengthened by effective copywriting.
Copywriters must create engaging, informative, and convincing text while retaining clarity and
brevity. Understanding the requirements and tastes of the target audience is critical for crafting a
message that will connect (Kapferer, 2008). Artificial intelligence opens up new opportunities in

56
various sectors, including scientific research, business applications, and everyday life. Creativity is
one of the most significant and, until recently, least understood of all intellectual talents among the
human dominating abilities (Wu et al., 2023). When used by expert hands, A.I. can boost human
creativity and makes copywriters' jobs more straightforward, quicker, and more efficient.

This chapter explores the impact of A.I. and Language Models in the broader marketing and
advertising area. The transformative power of A.I. revolutionized the world of marketing and
advertising and how customers relate to brands. (Campbell et al., 2020; Davenport et al., 2020)
For this reason, the investigation will also consider consumer behavior and how, in addition to
other societal shifts, the quick development of artificial intelligence has changed how companies
connect with their customers (Duarte & Neumaier, 2022). After covering these more broad but
fundamental topics, we will investigate the impact of generative A.I. on creating written content,
from journalism to copywriting. Subsequently, we will analyze the previous literature on the
perception of A.I. generated content, the distinction between human-generated content, and the
debate around it. The chapter will close by analyzing A.I.-generated content's ethical and legal
considerations.

4.2 A.I. and Language Models Use in Marketing and Advertising

Artificial intelligence (A.I.) represents a paradigm shift, not merely a technological improvement.
Its transformational impact is changing old frameworks and procedures as it pervades numerous
industries, from healthcare to finance. This revolution is evident in marketing, where A.I. is
improving operational efficiencies and reinventing the core methods that promote customer
engagement and brand loyalty (Campbell et al., 2020). Historically, Large Language Models
(LLMs) and natural language generation (NLG) primarily served as grammar systems, relying on
basic statistical models and rules (Elkins & Chun, 2020). NLG and LLM capabilities now go
beyond simple text generation. These models are modeled after human-like methods of translating
complicated ideas into comprehensible written or spoken language. They can comprehend context,
detect subtleties, and even respond creatively. NLG and LLMs, like computerized language
translators, bridge the gap between data and meaningful communication, adjusting to user inputs
and providing content that resonates with specific audiences.

57
As we anticipate the future of marketing, artificial intelligence's role is shifting from being
operationally helpful to being a critical strategic component, profoundly transforming the
landscape of customer-brand relationships (Elkins & Chun, 2020; Stone et al., 2020). The rapid
development of A.I., associated with cloud-based resources and platform-based business models,
leads to automated and interconnected business processes that have profound implications for
customers and stakeholders (Campbell et al., 2020). A.I. offers marketers substantial prospects for
new human-machine interaction, with applications spanning advertising, strategy, logistics, and
customer experience. This type of integration gives vital information into targeting the proper
consumers, successfully connecting with them, and undertaking return-on-investment analyses.
A.I. is rapidly powering various consumer-brand interactions, enhancing the customer experience.
It contributes to creating valuable, automated solutions to challenges that would otherwise
necessitate human intelligence, and it is critical in redefining the landscape of customer-brand
interactions to improve the whole customer journey (Campbell et al., 2020; Stone et al., 2020).
A.I.'s influence is evident in reshaping business models, optimizing sales processes, elevating
customer service, and deriving richer consumer insights (Davenport et al., 2020). It assists in
processes such as risk identification, consumer research, and tailoring operational tactics for
distinct customer segments. Moreover, A.I. is shaping marketing strategies, refining market
targeting, product positioning, and optimizing communication channels (Stone et al., 2020). Its
growing impact is also evident in pricing and competitive positioning. The advent of A.I. is
supported by the vast availability of data and the increasing affordability and speed of computing
power, offering insights beyond traditional statistical methods and reshaping the customer
experience to optimize the customer journey (Campbell et al., 2020).

Considering the diverse and highly distinct marketing sectors, it is evident that artificial
intelligence will have different aims, functions, and outcomes in each field. Chintalapati and
Pandey's (2022) analysis underscores A.I.'s transformative impact across various marketing
dimensions. In digital marketing, A.I. revolutionizes content creation, lead generation, customer
experience management, and social media engagement (Chintalapati & Pandey, 2022; Van Esch &
Black, 2021). Rabbi, Chimhundu, and Hassan (2021) emphasize A.I.'s role in fostering enduring
customer relationships and enhancing satisfaction.

58
In content marketing, A.I. plays a crucial role in content creation, management, and personalization
(Chintalapati & Pandey, 2022). Kose and Sert (2016) highlight A.I.'s ability to dynamically adapt
digital content based on user interactions, ensuring content relevance. Experiential marketing, a
rapidly evolving domain, is witnessing innovations like A.I.-driven voice interactions, virtual
reality, chatbots, and image recognition advancements (Chintalapati & Pandey, 2022). The
emergence of Marketing Technology (MARTECH) signifies A.I.'s profound influence on strategic
decisions in this area (Baltes, 2017). Market research, a foundational marketing element, is
leveraging A.I. to gain deeper consumer insights, effective segmentation, and actionable
intelligence (Davenport et al., 2020). Chintalapati and Pandey (2022) provide a comprehensive
framework addressing the ethical and security aspects of A.I. integrations. Ethical concerns, such
as A.I. potentially replacing human jobs, are prevalent (Dargham et al., 2022). However, A.I.'s role
in digital marketing is viewed as complementary to human expertise rather than a replacement
(Dargham et al., 2022; Wu et al., 2023). The objective is a harmonious synergy where digital
marketing benefits from A.I.'s precision and human creativity (Rabby et al., 2021).

Among the dominant human skills, creativity is one of the most important and, until recently, least
understood intellectual capabilities (Wu et al., 2023). With the aid of A.I., individuals can amplify
their creative potential, reaching results that would be either impossible or highly time-intensive to
humans alone. Since 2015, there has been a rapid increase in the demand for A.I.-related skills in
online job postings, including A.I.-enhanced creativity. Roles such as technology developers,
technology-product innovators, and product users exemplify the range of A.I. competencies, and
they must excel in A.I.-driven creativity (Alekseeva et al., 2020). Regardless of their skills or
specialties, individuals have ample opportunities to enhance their A.I. creativity (Wu et al., 2023).
Across all sectors and professions, there is a noticeable gap in A.I. expertise. While the demand is
particularly evident in the computer and information technology fields, it is not exclusive to them.
The professional services and financial sectors also highly need A.I. knowledge. Moreover,
industries like manufacturing, office operations, and agriculture are witnessing a significant uptick
in the relative demand for A.I. proficiency (Alekseeva et al., 2020). In summary, research indicates
that while A.I. can powerfully augment human creativity and abilities, it cannot fully replace them
(Wu et al., 2023).

59
In the dynamic world of contemporary marketing, social media is essential in shaping consumer
purchasing behaviors (Bond et al., 2010; Shaw, 2009; Zia et al., 2018). Marketers must connect
goods with consumer requirements, especially as customers become more outspoken about their
choices online. Brands may improve their image and reduce the likelihood of online criticism.
Traditional media, such as newspapers and television, were formerly dominant, but they are now
surpassed by digital strategies such as content marketing, SEO, and social media optimization
(Kiran, 2021; Kissel & Büttgen, 2015). A.I. supported the shifting from traditional marketing
tactics, such as direct or proxy emails and phone-based approaches, to more complex strategies that
leverage large amounts of data from the web, social media, and emails (Kiran, 2021). This data-
driven strategy, aided by artificial intelligence, enables more tailored, flexible, and interactive
advertising techniques. Marketers can now evaluate massive volumes of data from several
platforms in real-time, obtaining insights that improve campaign performance and expedite return
on investment (ROI). This efficiency improves advertising effectiveness and frees marketers' time
to focus on other essential tasks.

As marketing has evolved with A.I.'s integration, advertising has also transformed. Today's
advertising landscape extends beyond mere creative campaigns and selling tactics. Defined as a
paid communication medium with an identified initiator using mass media to resonate with
audiences, advertising aims to persuade the consumer to take some actions, usually to buy the
product (Moriarty et al., 2009). Nowadays, people are continually exposed to multiple classic mass
media, like television and newspapers, and new media platforms, like the Internet and social media
(Zia et al., 2018), and therefore get desensitized to classical advertising strategies (Baltes, 2017).
The challenge for brands is multifaceted: crafting content that captivates but stands out in an
oversaturated market and selecting media channels that incorporate novel and immersive
technological innovations (Duarte & Neumaier, 2022). In the face of current advertising issues,
artificial intelligence (A.I.) usage is wide and developing. A.I., for example, assists in creating
individualized promotions in computational and programmatic advertising. It accomplishes this by
automating ad scheduling, placement, and complicated media planning and procurement

60
procedures, maximizing ad reach and effectiveness (Huh & Malthouse, 2020). A.I. is
revolutionizing the advertising landscape by seamlessly integrating brand development strategies
with the impartation of tangible and perceived value. This integration is meticulously designed
with an overarching goal: to cultivate genuine consumer loyalty and establish a profound, lasting
connection between consumers and brands (Wijaya, 2012). The transformation we are witnessing
in the advertising sector is happening in collaboration, as the advent of innovative media platforms
is accelerating it. These platforms, with their advanced capabilities, enhance the diversity of
message delivery channels, ensure better content integration across mediums, and amplify the
overall impact of advertising campaigns (Kissel & Büttgen, 2015; Wijaya, 2012).

With the capabilities of A.I., advertisers can now pinpoint their target audience with unprecedented
precision, allowing for heightened personalization in their campaigns. This precision, in turn, leads
to more relevant and resonant advertising content. As a result, A.I. has emerged as an indispensable
tool in the advertiser's toolkit. It aids advertisers in delving deeper into customer behaviors,
extracting actionable insights, and amplifying key metrics such as customer targeting, lifetime
value, engagement, experience, and loyalty (Davenport et al., 2020).
The advancements in A.I. are more comprehensive than just data analysis. With state-of-the-art
technologies like OpenA.I.'s GPT-3, which has the capability to produce highly contextual and
relevant responses, the potential for creating A.I.-powered content is vast and largely untapped
(Duarte & Neumaier, 2022).

However, as with any rapidly evolving domain, the realm of A.I. in advertising research presents
its complexities. While the importance of A.I. in modern advertising is undeniable, there needs to
be comprehensive theoretical frameworks guiding its application. Various theories and models
have been proposed, but the industry still awaits a dominant, cohesive theory that can serve as a
guiding light. This gap underscores the pressing need for deeper, more expansive theoretical
exploration.

61
4.2 Artificial Intelligence and its Effects on Consumer Behavior Patterns
The marketing and advertising industries have seen considerable disruption since the rise of the
Internet, which has accelerated the need for adaptation and transformation. With the introduction of
A.I., the transformation was even more radical, and A.I. confirmed its transformative force. This
transformation impacted how the consumers see the brand, how it interacts with it, and where
(Duarte & Neumaier, 2022; Kissel & Büttgen, 2015; Wijaya, 2012). Especially evident through the
integration of chatbots, A.I. technologies have revolutionized advertisement delivery mechanisms
and fundamentally transformed the interaction dynamics between businesses and consumers
(Duarte & Neumaier, 2022).

A standout feature of A.I. in advertising is its ability to offer tailor-made personalization (Van Esch
& Stewart Black, 2021). This allows businesses to reach their target audience more accurately and
efficiently (Donahue & Hajizahed, 2019). A.I. relies heavily on data rather than standard human-
led strategies based on instinct or experience. This data-driven method, typical of machine
intelligence, turns vast amounts of information into strategies that directly influence consumer
actions (Haenlein & Kaplan, 2019). Building on the foundation of A.I.'s transformative impact on
advertising and its new capacity for personalization, it is crucial to delve deeper into one of its
most influential aspects: the shift to conversational mediums. Van Esch and Stewart Black (2021)
underscore A.I.'s impact on digital marketing personalization. Their work highlights that, with
enhanced data collection and processing capabilities, marketers are now in a position to serve
consumers with highly individualized advertisements in real time. They also envision a future
where A.I. facilitates pinpointed advertising based on a consumer's location within a store, raising
the bar for personalization.
Further emphasizing the gravitation towards personalization, advertisers are no longer focusing on
broad strategies but are increasingly monitoring individual online behaviors. Such techniques have
culminated in online behavioral advertising (OBA). As Boerman et al. (2017, p. 364) describe,
OBA monitors people's online behavior and uses the collected information to show people
individually targeted advertisements. These techniques are viewed as integral to the future of
advertising, creating avenues to connect with specific audiences in unprecedented ways.

62
However, Boerman et al. (2017) argue that the type and quantum of personal data leveraged govern
the nature of personalization in OBA. The spectrum of data spans browsing and search histories,
while its quantity delves into the depth and mix of such information. A distinctive characteristic of
OBA, as pointed out by Boerman et al. (2017), is the covert way personal data is collected, raising
ethical concerns as consumers often remain oblivious to the tactics advertisers deploy.
Mogaji et al. (2020) delve into the potential of such data, indicating that corporations can utilize
online information to prepare and personalize advertisements that resonate emotionally with
consumers. They set that ads can be crafted by selectively merging creative elements with the aid
of A.I. (Mogaji et al., 2020). These resulting ads, tailored to the individual and emotionally
appealing, are disseminated digitally, illustrating the power and implications of programmatic
advertising. Programmatic advertising pertains to the automated buying of ads, where algorithms
and data analytics determine the ads' relevance and associated costs. (Mogaji et al., 2020)

As a primary A.I. tool, Chatbots have played a considerable role in steering this evolution. They
have transformed advertising from a general, mass communication approach to a more nuanced,
conversational style, bridging the gap between businesses and consumers in a previously
unimaginable manner (Kim et al., 2022). This transition to conversational advertising means that
today's social media users expect and anticipate a multimedia-rich content journey, with
meaningful brand interactions prioritizing generic ad slogans (Duarte & Neumaier, 2022).However,
it is essential to understand that not all personalization approaches generate universal success.
Two studies analyzed by Kim et al. (2022) found that promotion-focused consumers favor
personalized chatbot advertising, seeing more usefulness in sharing personal information.
However, prevention-focused consumers view these ads less favorably due to a higher perceived
risk. Those with high privacy concerns dislike such personalized ads, no matter their regulatory
focus. The balance of perceived risk and benefit mediates the impact of ad personalization,
regulatory focus, and privacy concerns. (Kim et al., 2022) Duarte & Neumaier (2022) emphasize
another potential challenge, noting that the effectiveness of conversational advertising may vary
across demographic segments. For instance, younger consumers (aged 18-34) have demonstrated a
higher receptiveness to chatbot-assisted shopping, indicating a generational divide in A.I.

63
acceptance. Such insights require brands to discern their target demographics and gauge their
receptiveness to A.I.-driven interactions.

Focusing on targeting, the prowess of A.I. in analyzing vast swathes of data and discerning patterns
can significantly enhance a company's grasp of its target audience and their preferences. As Stone
et al. (2020) indicate, this deepened understanding paves the way for more resonant marketing
campaigns and finely tuned product propositions, holding the promise of driving revenue
growth.Targeting appears as a critical frontier where A.I. showcases its transformative potential.
A.I. applications, notably chatbots, do not just personalize but enable marketers to finetune their
approaches, focusing on specific consumer demographics with unprecedented precision. As noted
by Liu (2020), the evolution of targeting technology has cemented its function as a turning point in
contemporary marketing strategies. This refined targeting raises the hit rate of advertising and
optimizes the allocation of marketing resources, ensuring each campaign garners maximum return
on investment. However, as nuanced as A.I.'s capabilities are in this domain, the process is only
partially automated. While fundamentally a mechanical operation, segmenting the market does
benefit from A.I.'s data processing prowess. However, deciding which segment to zero in on
requires more than algorithmic determinations. Rust et al. (2021) stress the indispensability of
human expertise in this equation, as humans's knowledge, strategic discretion, and intuition remain
irreplaceable, ensuring A.I.'s analytical findings are aptly contextualized and actioned upon.
Brands must employ these A.I. technologies intelligently, realizing that their performance varies
across customer segments and from previous communication channels. They should be seen as a
component of a broader advertising strategy, as emphasized by Duarte and Neumaier (2022). In the
ever-evolving business landscape, continually refining and adapting organizational value chains
cannot be exaggerated. A.I.'s digital marketing enhancement presents a refined approach to
customer engagement, ensuring businesses connect with their audience with impeccable timing
(Ransbotham et al., 2017). Rabby et al. (2021) further illustrate the benefits of incorporating A.I.
into digital marketing, noting the potential to foster lasting relationships and elevate customer
satisfaction. By leveraging tools like chatbots, big data, and cognitive technologies, retailers see
remarkable results; those using A.I. outperform traditional methods by a factor of five. (Rabby et
al., 2021)

64
Trust issues can arise as digital marketing shifts consumer expectations towards more tailored
interactions. Many modern consumers demand transparency and information before commitment,
emphasizing the need for businesses to be forthright and ethical in their A.I.-driven approaches
(Singh et al., 2019). Kim et al. (2022) also pointed out the reticence of those predisposed to more
significant privacy concerns towards using chatbots. Due to these A.I.-driven strategies, an
increasingly discerning clientele is emerging. Many businesses nowadays dedicate significant
effort to in-depth research before making purchase decisions. This behavior is emphasized by
Yussaivi et al. (2019, as quoted by Rabby et al., 2021), underscoring the complex interplay
between A.I. innovations and evolving consumer habits.

In conclusion, A.I.'s impact on consumer behavior is significant and evolving. As technology and
targeting capabilities advance, A.I. will likely become increasingly prevalent in consumer-business
interactions. Despite this, it is crucial to recognize the complexity of online chatting and the need
for authentic interactions. A.I., especially in the form of chatbots, can be a powerful tool. However,
it must be carefully incorporated into broader advertising and communication strategies to
influence consumer behavior, as Duarte & and Neumaier (2022) highlighted.

4.3 The Revolutionary Influence of Artificial Intelligence on Written Content Creation

Advancements in Artificial Intelligence have catalyzed transformative changes in both the business
and creative realms, significantly influencing diverse categories of written content, including
journalism and creative writing. Over the past decades, improvements in A.I. have extended its
capabilities far beyond grammar and spell-checking functions (Clark et al., 2018). In its early
stages, natural language generation was primarily confined to strict grammar systems, rudimentary
statistical models, and comprehensive rule-based guidelines. These initial systems had a
constrained capacity, often resulting in garbled or limited-scope outputs. However, the introduction
of large-scale statistical models, exemplified by GPT-3.5, has revolutionized the domain. Unlike its
predecessors, GPT-3.5 captures the nuances of language through extensive exposure and learning,
rather than adhering to rigid predefined rules (Elkins & Chun, 2020). A.I. models such as GPT-3,
GPT-3.5, and GPT-4 have been developed with extensive cognitive capabilities, allowing them to

65
assist users in a wide range of tasks rather than being confined to a singular objective (Lee et al.,
2023). This versatility empowers them to undertake roles that, prior to the advent of such A.I.
systems, were exclusively human domains, including the generation of novel creative content. Due
to advancements in artificial intelligence, the era of computer-generated content is now upon us
(Dargham et al., 2022).

The influence of generative artificial inteligience on content creation is undeniably significant.

Prestigious journalistic platforms, including The New York Times, The Washington Post, and
Business Insider, have already incorporated A.I.-crafted narratives into their content strategies
(Dargham et al., 2022). A.I.-driven softwares may assist authors in editing, structuring, and
refining their work, as evidenced by word processors, grammar and spell checks, version control,
and even language or style analysis tools (Clark et al., 2018). Furthermore, they can be helpful in
functions such as brainstorming and outlining an article, helping the author to utilize their
creativity at best. Overall, A.I. makes the writer's job more efficient, faster, and more satisfactory
(Wu et al., 2021).

A.I. can represent an outstanding tool, not just for journalists but for all writers. While crafting
comprehensive narratives has long been the exclusive domain of human intelligence,
advancements in natural language generation and A.I. have targeted this attribute (Klein et al.,
1973; Turner, 1993). With its nuances of coherent narratives, distinct authorial tones, and the
ultimate goal of telling a story that can generate feelings in the reader, creative writing and story
creation is a unique challenge for generative A.I. (Harper, 2010). What makes it a challenge is the
length of the tale plots, several character identities, and the author's distinct tone (Ippolito et al.,
2022). However, based on the existing literature, we can affirm that the breakthroughs in natural
language models are also revolutionizing the discipline of creative writing (Ippolito et al., 2022;
Klein et al., 1973; Turner, 1993). While initial evaluations of this A.I. software have predominantly
been by amateur writers (Clark et al., 2018), insights from seasoned creative writers can offer more
profound, more nuanced feedback, illuminating potential integrations of A.I. within traditional
writing processes (Ippolito et al., 2022).

66
Another field where the use of A.I. at the center of a vigorous debate is academic writing or any
writing related to education. The research from Fitria (2023) proves that ChatGPT, utilizing
GPT-3.5, can correctly write English essays. The bot provided detailed responses when the
researcher asked ChatGPT for help with an English assignment. ChatGPT has proven to be capable
of handling subjects, including writing descriptive texts, sharing personal experiences, and
predicting outcomes in several experiments. It used both active and passive voices acceptable for
the tense and paid close attention to the structure, including introductions, primary material, and
conclusions. However, more research into the grammatical correctness of writings created using
ChatGPT is still necessary (Fitria, 2023). ChatGPT can provide well-written student essays,
summaries of research papers, answers to medical test questions that are sufficient to pass, and
helpful computer codes (Van Dis et al., 2023).

Dergaa et al. (2023) highlighted how ChatGPT has several advantages for academic research,
including fast processing of massive amounts of text input, automating processes like scanning and
extracting information from academic publications and producing unbiased summaries of lengthy
works. It can even propose prospective study ideas based on input subjects. These features make it
a priceless resource for scholars, particularly those new to a topic or looking for project ideas.
Although ChatGPT has much potential, it is crucial to utilize it wisely, considering its restrictions,
appropriately acknowledging sources, and addressing delicate subjects carefully (Dergaa et al.,
2023), especially as it is hard for the reader to identify if a text has been generated by A.I. or
written by a human (Van Dis et al., 2023). Researchers should use it with other techniques and
keep up with NLP developments (Dergaa et al., 2023).

Regarding copywriting, Artificial Intelligence has rapidly claimed its utility, becoming an
instrumental catalyst for several intricate tasks, from content origination to grammar checking.
(Clark et al., 2018; Wu et al., 2023) Wu et al. (2023) indicate that contemporary advancements
empower A.I. to perform complex assignments such as assimilating incoming data and then
autonomously crafting A.I.-generated content tailored for the digital marketplace.However, it is
crucial to highlight a conspicuous void in the existing literature. There needs to be more in-depth
research specifically scrutinizing the tangible impact of A.I. within the specialized niche of

67
copywriting. This absence underscores a potential avenue for further exploration, suggesting that
while A.I.'s overarching influence on digital content is being recognized, its granular effects on the
art and science of copywriting remain largely uncharted.

4.4 Distinction Between Human-Generated and Computer-Generated Content

The distinction between human-created and computer-generated content increasingly diminishes
the more evolved machines become. Is it A.I.-generated content if a human author writes his text
but uses an A.I. to correct the grammar and make it more readable? Which percentage of human
control, input, or inventiveness makes content human generated?

Racine (2018) defines A.I.-generated content (AIGC) as works in which human authorial influence
is largely absent. A.I. systems transition from mere tools to central creators, autonomously making
creative decisions rooted in their internal algorithms. This development in A.I. role is evident in
endeavors like "The Next Rembrandt", where the substantive creative determinations stem
primarily from the A.I., even within human-established parameters. Wu et al. (2023) traced AIGC's
development through three key stages. In the initial phase, primitive programming was used to
create content, such as the Iliac Suite music in 1957 and Eliza, an early interactive robot. The
second phase was defined by the growth of massive databases and powerful computers, which
resulted in contributions such as the novel 'The Road' and Microsoft's real-time translation system.
However, beginning in 2010, the third and most revolutionary phase saw tremendous
developments, highlighted by the debut of ChatGPT in 2022 (Wu et al., 2021). The capacity of this
chatbot to understand and write human-like prose demonstrates AIGC's limitless potential.

Further characterizing AIGC, it may be broadly classified as producing text, photos, audio, and
videos:
• Text Generation: The text-based components of AIGC include organized, creative, and
dialogue-focused writing (Wu et al., 2023). In this category are included chatbots, whose
primary goal is to improve communication between humans and machines (Zhang et al.,
2023). While structured writing curates material from certain data sets, creative writing,
appropriate for promotional content and blogs, necessitates a one-of-a-kind personal touch.

68
 Chatbots that simulate customer service encounters benefit from dialogue scripting (Wu et
al., 2023).
• Image Generation: AIGC permits image changes and additions in particular directions (Fan
et al., 2020). This technique simplifies picture editing, making creating even without prior
knowledge possible. A GC can also produce pictures or turn 2D visuals into 3D
visualizations (Pollefeys & Gool, 2002).
• Audio Generation: AIGC's aural capabilities range from text-to-speech synthesis to voice
reproduction. Text-to-speech conversion, prevalent in robotics and voice broadcasting, is
reaching new heights and may soon advance to emotive speech synthesis. Voice cloning
imitates individual voices from provided samples, which might be helpful in domains such as
intelligent dubbing (Wu et al., 2023).
• Video Generation: AIGC is largely used in video editing to create promotional content and
trailers (Texler et al., 2020). This approach is similar to picture generation in that each video
frame is processed independently. A GC can create attractive promotional films by
integrating different A.I. algorithms, marking a significant shift in video content development
and promotion (Wu et al., 2023).

4.5 Perception of A.I. Generated Contents

Since A.I. is now capable of using language to perform tasks that were previously extraordinarily
challenging or perhaps impossible (Elkins & Chun, 2020), it has become increasingly more
challenging to understand what human-generated content and A.I. generated content. According to
Ballardini et al. (2019), A.I. systems, powered by massive data sets and advanced algorithms,
replicate intelligent human behaviors, and this capability calls into question our understanding of
the human creativity-machine autonomy continuum. Racine (2018) delves more into the perceptual
issues posed by AIGC. He distinguishes between two categories of content: collaborative human-
computer creations and purely A.I.-generated works. In collaborations, computers typically support
human creators, but occasionally, their roles become ambiguous, primarily when A.I. heavily
influences the final piece. This blurring raises questions about the true extent of human
contribution and its impact on perceived originality. Even though humans might set initial
parameters for purely A.I.-generated content, the core creativity is governed by A.I. protocols. The

69
striking similarity of some A.I. creations to human works sparks debates on their authenticity and
emotional depth. Critics contend that A.I.-produced content might never capture the emotional
richness of human-made works. Further, Racine (2018) highlights a tension between content
perception and its creation process. The current Copyright Act emphasizes the latter, leading to
questions about a programmer's contribution to content "originality." While rule-based systems can
produce unpredictable outcomes despite a structured foundation, neural networks craft much of
their algorithms independently, distancing the end product from human input. While some argue
whether A.I.-generated material can compete with human-produced works, skeptics and
proponents cope with this changing landscape (Latar, 2015).

Understanding this topic requires a journey into the world of perception. Perception is typically
considered a conscious process in which people are vividly aware of and can describe their feelings
(Kihlstrom et al., 1992). On the other hand, modern research has dug into nonconscious perception,
separating it from conscious perception with terms like explicit and implicit perception (Kihlstrom
et al., 1992). Explicit perception refers to conscious knowledge of external or internal stimulations,
whereas implicit perception refers to behavioral, cognitive, or emotional adjustments caused by
unobserved contextual events (Kihlstrom et al., 1992).

There are occasions when these two perceptions converge, but they might also part ways, notably
in scenarios tinged with prejudice or stereotypes (Greenwald & Banaji, 1995; Greenwald et al.,
1998). Such variations stem from societal norms that might frown upon explicit prejudice but are
discernible through subtle gestures (Castelli et al., 2001; Maasset al., 2000) that indicate implicit
perceptions. In deciphering this, Sloman's (1996) dual systems of reasoning model emerges as
enlightening. This model posits one system centered on rapid, conscious thought, while its
counterpart delves into slower, nonconscious information processing. Notably, overlooked cues
mold implicit perceptions, whereas explicit ones lean on verbal cues (Rydell et al., 2006).
Surprisingly, this paradox is not confined to prejudice. Engaging with A.I.-generated information
may cause the quick system to develop a conscious evaluation while the slow system forms an
unconscious one. Whether explicit or implicit, A.I. content perception emerges autonomously (Wu
et al., 2020).

70
The concept introduced by Sundar (2008, as quoted by Wu et al., 2020) is central to this debate.
This point of view considers nonhuman agents to be more objective than humans. Graefe Haim et
al. (2018) further support this stance, claiming that machine-authored information has higher
credibility and demonstrates more profound knowledge than human-written content. Clerwall
(2014) investigated people's perceptions of news trustworthiness and readability. When authorship
was not mentioned, the results indicated a modest difference in rating text credibility and
readability. A.I.-generated news looked complete and neutral but less entertaining than human-
written items. Using software to reveal if a machine or person created a text is already possible, but
advancing A.I. technology and intelligent prompts may outperform existing detection techniques
(Van Dis et al., 2023).

4.6 Ethical and Law Considerations about A.I.-generated Content

ChatGPT and other LLMs are undeniably helpful tools for conversational and writing jobs, helping
to improve the effectiveness and correctness of the desired output (Castellanos-Gomez, 2023).
LLMs have been developed for years, but recently, they have become considerably more potent
than previously because of constant improvements in the quality and amount of data sets and
advanced techniques for calibrating these models with human feedback (Van Dis et al., 2023).

On the other hand, questions have been raised about the opposing sides of A.I. concerns with A.I.
language models, in particular with Chat-GPT, include ethical, copyright, transparency, and legal
difficulties, as well as the dangers of prejudice, plagiarism, a lack of creativity, false material that
might cause hallucinations, a lack of expertise, wrong citations, cybersecurity problems, and
infodemic dangers (Sallam, 2023). Another main point of the discussion is the potential bias based
on the datasets used in ChatGPT training, which may restrict its capabilities and lead to factual
mistakes yet alarmingly seem to be scientifically reasonable, a phenomenon known as hallucination
(Deng & Lin, 2023). Castellanos-Gómez (2023) highlight another potential use of A.I. that is at the
center of an intense debate: the possible exploitation of ChatGPT and other A.I. language models to
support academic paper and also mislead readers about the caliber of the study. He advocates for
including a paragraph outlining the precise usage of A.I. language models in scientific journals. He
advises that the scientific community create ethical standards and best practices (Castellanos-

71
Gómez, 2023). The produced text is the researcher's property and may be used in their article
without restrictions under OpenAI's shared publication policy. OpenAI, however, advises that the
article should contain some wording that discloses the usage of an A.I. language model.
Castellanos-Gómez enthusiastically endorses this idea and offers an illustration of an
acknowledgment that might be used to explain how ChatGPT was used to create the material. While
A.I. language models can help with various elements of writing scientific articles, he notes that it is
still crucial to give credit where credit is due (Castellanos-Gómez, 2023).

The capabilities and controversies of ChatGPT in producing academic publications are examined by
Anderson et al. in 2023. They review the advantages and disadvantages of employing A.I. to
generate sports and exercise medicine manuscripts, potential issues with A.I. based on natural
language models, and whether A.I. should write academic articles. The authors asked ChatGPT to
produce two articles on the advantages and disadvantages of employing A.I. in scientific
publications and sports medicine to demonstrate ChatGPT's capabilities. The produced essays were
created in seconds instead of the 45 days it took a group of seasoned writers to create the present
editorial (Anderson et al., 2023).

Furthermore, the emergence of artificial intelligence has sparked lengthy disputes over the rights
bestowed upon its creations. Ballardini et al. (2019) identify three primary schools of thought in
these disputes. The 'Revolutionary School' contends that A.I. should be granted human-like rights.
This viewpoint delves further into the ethical quagmire, wondering whether computers should be
treated similarly to living creatures regarding rights and obligations (Ballardini et al., 2019). In
contrast, the 'Romantic School' takes a more conventional perspective, arguing that only people
should be granted content rights. While this approach values human creativity, it may neglect the
growing significance of artificial intelligence in creative processes (Ballardini et al., 2019). The
'Modernist School' bridges these two viewpoints. This school of thinking proposes a more nuanced
approach in which people at the helm of A.I.'s creative products are granted rights. The conundrum
here is establishing the limits of human engagement and how to fairly give credit, especially when
A.I. plays an increasingly unexpected and dominant role in production (Ballardini et al., 2019).
Aside from these schools of thinking, there is an increasing debate over the economic ramifications
of A.I. inventions, particularly the equal compensation of human producers (Hancock et al., 2020).

72
The demand for transparency in A.I. characteristics is a new topic in A.I. discourse (Van Dis et al.,
2023). While certain technologies, such as autocorrect, have become practically invisible in our
daily interactions, others, such as Deepl (DeepL Nuova Svolta Nella Qualità Della Traduzione
Basata Sull'Intelligenza Artificiale, 2020) or Grammarly (How We Use A.I. to Enhance Your
Writing | Grammarly Spotlight, 2019), proudly display their A.I. background. Legislation, such as
California's B.O.T. Act of 2018, aims to increase transparency in this area (Hancock et al., 2020).
Such efforts, however, raise several issues, ranging from potential human rights violations to
unknowingly burdening A.I. users. Wu et al. (2023) raise concerns about how artificial intelligence
systems handle sensitive personal data, emphasizing a rising wave of privacy dangers. The privacy
issues surrounding ChatGPT and other language models are complex and cover various topics,
including data protection strategies and ethical considerations. Brown et al. (2022) point out the
drawbacks of well-liked data protection strategies like data sanitization and differential privacy,
contending that these approaches cannot provide language models with a general and meaningful
notion of privacy. They propose that text data generated explicitly for public usage should be used
to train language models. Similarly, Shahriar and Hayawi (2023) emphasize ChatGPT's potential
value across a variety of sectors while also carefully examining the ethical and privacy issues raised
by its use. They underline the significance of comprehending ChatGPT's growing potential while
acknowledging its limits. However, Mosca and Such (2022) concentrate on creating ELVIRA, an
explainable personal assistant created to improve the sharing policy for collectively owned content.
They contend that this strategy may support multiuser privacy more successfully than other
approaches in the literature. Last but not least, Silva et al. (2020) suggest using Named Entity
Recognition to recognize, track, and validate Personally Identifiable Information, while also
exploring the potential of this strategy as a Privacy Enhancing Technology and the risks and effects
connected (Brown et al., 2022; Shahriar & Hayawi, 2023; Mosca & Such, 2022; Silva et al., 2020).
Wu et al.'s work (2023) also addresses concerns such as A.I.'s potential difficulties with subjective
material and plagiarism, emphasizing the critical need for human oversight. De Bruyn et al. (2020)
highlight A.I.'s ability to understand complex data matrices and autonomously construct models in
their study. However, this natural power can occasionally have unanticipated repercussions.
Meanwhile, Wahle et al. (2023) express concern about over-reliance on A.I. models for content
development. They highlight potential hazards like errors and citation issues. Given the changeable
nature of regulatory frameworks governing A.I. content, they ardently argue for openness and a
keen understanding of the surrounding environment, particularly in sensitive industries such as

73
healthcare. The intersection of A.I., ethics, and law is fraught with obstacles and opportunities.
Balancing A.I.'s inventiveness with ethical considerations and legal protections is complex.

4.7 Conclusion
As deeply analyzed in this chapter, the evolution of artificial intelligence undeniably presents a
transformative force in content creation. The pioneering innovations by algorithms like ChatGPT
testify to the commercial viability and dynamic performance of A.I.-Generated Content (AIGC)
technology. With vast volumes of high-quality data and the prowess of high-performance hardware,
many methods for these expansive models have emerged, reshaping how content is created and how
it influences and interacts with its audiences (Wu et al., 2023).

However, as Korteling et al. (2021) state, while A.I. capabilities are immense, the adaptability,
emotional intelligence, and intricate understanding inherent to human intelligence remain
unparalleled. This organic blend of creativity and emotional depth from our evolutionary history
distinguishes us as humans. The prospect, therefore, is not a choice between human intelligence and
A.I. but rather a harmonious synergy of both. Collaborations between the two might unlock
potential hitherto unimagined, granted that the A.I. output remains accurate and trustworthy.
Nevertheless, as the horizons of A.I. expand, its influence begins to permeate deeper layers of
human interaction. Hancock et al. (2020) elucidate the transformative power of A.I.-mediated
communication (AI-MC). A.I. innovations like Gmail's "smart reply" and "smart compose" are not
mere tools but potentially culture-shifting entities. These systems often prioritize positive prompts;
they might unknowingly direct human communication in an increasingly affirmative direction.
( Hancock et al., 2020). While optimism is undoubtedly commendable, the inadvertent emphasis on
specific linguistic patterns or terminologies, especially those emerging from a constrained semantic
pool, risks diminishing the rich tapestry of human communication. The optimization, therefore, lies
in ensuring that while we leverage the efficiencies of A.I., we retain the depth and diversity of
human expression. (Hancock et al., 2020)

In conclusion, the future of AIGC is bright but challenging. A conscientious, ethical, and
collaborative approach will be essential to utilizing A.I.'s promise without overshadowing human
creation's inherent complexity and nuance as it continues to influence the digital environment. The
choreography of the dance between people and robots in content production, which is still in its

74
early stages, will determine how innovation, expression, and interactivity are portrayed in the
coming years. Although the numerous effects of A.I. on different types of written material have
been intensely investigated in this chapter (Chintalapati & Pandey, 2022; Dargham et al., 2022;
Duarte & Neumaier, 2022; Wu et al., 2023), there is still a significant void in the research about its
specific impact on copywriting. A noticeable lack of in-depth studies devoted to comprehending the
subtleties of artificial intelligence's involvement in copywriting persists despite the abundance of
studies analyzing the effects of A.I. on content creation from creative writing (Dargham et al., 2022;
Ippolito et al., 2022; Klein et al., 1973; Turner, 1993) to academic writing (Brown et al., 2022;
Shahriar & Hayawi, 2023; Mosca & Such, 2022; Silva et al., 2020). Given the crucial role that
copywriting plays in influencing customer behavior and brand perception, this gap stands out as
very troubling. The need for compelling and engaging content has never been higher as the digital
world continues to change. However, it still needs to be discovered how much A.I. can do to help
satisfy this need without sacrificing the inventiveness, persuasiveness, and authenticity that
distinguish good copy. Future studies will have a compelling chance to explore this understudied
area and clarify how A.I. may be used to complement, rather than replace, the art of copywriting.

75
II. Empirical Study

5. Descriptive Memory of the Internship Report

5.1 What is legal technology?

As abundantly described in the literature review, every industry has been deeply impacted by the
revolution going on because of technological innovations, especially in the past decades. The legal
profession, often considered conservative and resistant to change, has not been immune to these
changes. The legal profession's need to update and adapt to the twenty-first century has become
increasingly apparent.

Webb (2020) defined legal technology as reshaping legal work and the organization and governance
of legal practice. It is widely accepted that Legal technology is a new idea that aims to integrate
technology tools into the legal system (Pasquale & Cockfield, 2019). Often referred to as legal
tech, legal technology is an interdisciplinary field that blends the fundamental principles of the legal
sciences with cutting-edge areas like computer technology, artificial intelligence, and data analysis
(Mania, 2022). Due to this convergence, a new class of services that uses technology breakthroughs
to improve, expedite, and reinvent the delivery of legal services has emerged. Salmerón-Manzano
(2021) explains how legal tech is the use of new technologies applied to the world of law to carry
out tasks that, until recently, were performed manually by lawyers. Mal'ko (2022) also highlighted
how legal technology allows the implementation of strategic tasks. Furthermore, legal tech can refer
to digital solutions that address more general legal issues, such as data protection, software, and
platforms created specifically for legal practitioners; the debates over the General Data Protection
Regulation's (GDPR) implications for speech data have brought this to light (Nautsch et al., 2019).

As discussed by Kerikmäe et al. (2018), the business model of law firms is changing, particularly
for services that entail data processing. A wide range of technology solutions, each with specific
objectives and growth curves, are driving this revolution. The advancements in legal technology are
changing the landscape of legal services, from automating repetitive operations to using artificial
intelligence for legal research and forecasts. These developments increase the accuracy and

76
efficiency of legal procedures while also democratizing access to legal information, improving the
accessibility and transparency of the judicial system.

To summarize, Webb (2020) defines legal technology as incorporating technology into law practice.
This integration encompasses many tools and applications, including software for efficient legal
document management, internet platforms connecting attorneys with clients, and analytical tools
forecasting legal case outcomes. By utilizing these technologies, lawyers and legal professionals
may dramatically improve their efficiency and production. Furthermore, legal technology
democratizes the legal system, making it more accessible and affordable to individuals and
corporations. As Kerikmäe et al. (2018) point out, these technology solutions influence both
attorneys and clients on several levels. Legal technology is evolving rapidly, with new
developments occurring regularly. Contract management software, e-discovery tools, and platforms
that provide online legal services are common examples.

5.2 Introduction to Sweet Legal Tech

Sweet Legal Tech (SLT) is a pioneering Italian startup specializing in technology integration and
consultancy within the legal sector. Sweet Legal Tech is driven by a passion for replacing
antiquated methods and harnessing the full potential of modern technology to deliver superior
outcomes for clients.

Established in 2018, Sweet Legal Tech initially emerged as a think-tank committed to the
technological education of legal professionals. Over the years, they have meticulously mapped,
examined, and experimented with many Legal Tech tools and have carved their niche by advocating
innovation in the legal field. Their educational initiatives, aimed at training lawyers with
technology, increased in Italy and across Europe. By the end of 2019, something had changed
already: participants from SLT's seminars began to perceive the company as an authoritative figure
in the discipline, seeking guidance for digital transformation endeavors. Consequently, from 2020
onwards, supported by a bigger team composed of two-thirds of lawyers, SLT embarked on a
journey to provide continuous consultancy and technology integration services.

77
SLT's partnership portfolio currently boasts collaborations with esteemed legal tech giants such as
Thomson Reuters, Luminance, OneTrust, Bigle Legal, Ironclad, and DocuSign. Furthermore, they
are privileged to be a part of renowned EU entrepreneurial networks like ELTA (The European
Legal Tech Association) in Italy. Complementing their industry affiliations, SLT's top-tier
management holds esteemed academic positions, teaching the "Legal Tech and Coding for
Lawyers" postgraduate course at the University of Milan and the "Legal Tech" module at the LL.M.
in Law of Internet Technology at Bocconi University. Moreover, Sweet Legal Tech offers
innovative services such as scouting and benchmarking analysis of legal tech solutions, integration
of Contract Lifecycle Management into legal departments, assistance with AI tools for contract
review, and support with e-signature workflows. A technical team with a background in law and a
specialization in technology analyzes and understands each client's problems and offers them a
solution, finding the best tool for their needs and helping them integrate it into their workflow to
make it more efficient.

Sweet Legal Tech's mission is to "help lawyers be lawyers" and not spend their time on tasks that
could be automatized and optimized, providing a better workflow and, consequently, a better work-
life balance. Since legal technology is still a recent innovation in the legal world, which, as pointed
out before, is a world strongly tied to traditions, a fundamental part of Sweet Legal Tech's mission
has always been to explain these tools educationally. A prime example of their dedication is the
"Legal Tech Index" they have curated, encompassing over 250 legal tech firms, solutions, and
software. More than just a listing, this index highlights a select group of tools, termed the "magic
circle," which have been thoroughly evaluated and endorsed by SLT's experts for their outstanding
quality. Additionally, Sweet Legal Tech provides unmatched Benchmarking Reports with in-depth
analyses and contrasts of various legal tech solutions inside specialized domains. This report
facilitates legal professionals in making informed choices by weighing the features, functionalities,
pricing, advantages, and drawbacks of the myriad options available in the legal tech arena. With
resources like the Legal Tech Map and Benchmarking Reports, Sweet Legal Tech stands as an
affirmed brand for legal professionals navigating the evolving landscape of legal technology.

78
5.2 The Structure of the Company
Sweet Legal Tech is a startup with limited resources and staff, therefore, it is not structured with
defined departments but thrives on a collaborative attitude and support. Every employee has a
substantial role in the growth of the business, which promotes a culture of cooperation and shared
goals. Nevertheless, it is possible to identify particular positions inside the company. The majority
of the staff concentrates on offering assistance and support to new and old clients, researching and
testing various legal tech tools, and integrating them into clients' working environments.
Additionally, it guarantees that Sweet Legal Tech's services are up-to-date and tailored to customers'
demands. A small sales and marketing staff in charge of increasing Sweet Legal Tech's influence
and reach complements this, but other people in the company are also expected to take care of these
functions. For example, before my internship, the company's CEO directly curated the
communication on social media. Furthermore, it is essential to highlight that the entire staff of
Sweet Legal Tech is hired as freelance; therefore, they have multiple projects to devote their
attention to.

5.3 Employer Branding Strategies of Sweet Legal Tech

The employer branding methods utilized by Sweet Legal Tech reflect this strategic positioning.
Analyzing their strategy reveals a fusion of conventional principles and contemporary sensibility to
attract and keep the best legal talent.
These are some of the main points to consider about their employer branding.
• Emphasis on Professional Advancement: Sweet Legal Tech highlights on its website the
opportunity for professionals to work with cutting-edge technologies shaping the legal
profession's future. Furthermore, they highlight that the entire team comprises young, brilliant
professionals with particular attention to innovation and technology development. By
emphasizing this, they target those who aspire to be at the forefront of their field, attracting
individuals looking for a job and a platform for professional growth and innovation who want
to change the legal field.
• Promotion of Direct Impact: Instead of hiding it, on their website, they mention the small
number of people that compose the company by underscoring the difference between being "1
in 1,000" in a larger organization versus playing a direct, impactful role at Sweet Legal Tech.

79
 In this way, the company positions itself as a place where individual contributions are valued
and recognized. This strategy appeals to professionals seeking more hands-on experience and
wanting to see their efforts' direct results.
• Continuous Learning and Curiosity: Sweet Legal Tech's dedication to training reflects its
commitment to developing its workforce. However, what distinguishes them is how they
foster inquiry. SLT promotes a culture of continuous learning and innovation by encouraging
curiosity and ensuring that passionate ideas are constantly acknowledged, making it an
appealing option for people who value personal and professional growth.
• Balanced Work Culture: The debate between remote and in-office work is constant in the
post-pandemic work environment. Sweet Legal Tech's approach to this is both pragmatic and
forward-thinking. By promoting a blend of remote and in-office work, they acknowledge the
benefits of flexible work schedules while emphasizing the importance of face-to-face
interactions. Furthermore, they empathize that freelancing means dividing your time over
multiple businesses. Therefore, being in the office from 9 to 5 every weekday would be
impossible and unproductive. This strategy not only caters to the evolving preferences of
modern professionals but also underscores SLT's belief in the value of direct collaboration and
personal responsibility.
In conclusion, Sweet Legal Tech's employer branding strategies are a testament to its commitment
to innovation in the legal tech domain and the holistic development of its employees. Their
strategies reflect a deep understanding of modern professional aspirations, making them a
compelling choice for those seeking a dynamic and fulfilling career in legal technology.

5.4 Initial Contact and Onboarding

My journey with Sweet Legal Tech began in September 2022. I was drawn to an advertisement on
LinkedIn, which prompted me to explore Sweet Legal Tech's official website. Intrigued by their
mission and the potential role, I submitted my CV, a motivation letter, and a portfolio through their
dedicated job application section. They were looking for a working student fluent in English and
Italian, as the social media pages and client contact are in English, but the company is based in
Italy. The rigorous selection process reflected the company's commitment to excellence and search
for individuals who could resonate with its vision. I had the first acquaintance interview, where I

80
was asked to craft an appealing case study based on their latest client experience that could be used
on their LinkedIn page, and a follow-up interview to discuss my production. After navigating this
meticulous process, I began my internship on October 3, 2022, at Sweet Legal Tech's central office
in Milan. The terms of my contract stipulated a six-month engagement, requiring a commitment of
40 hours per week with flexible timing, as during the last months of my internship, I had to attend
classes at university. Throughout this journey, I was fortunate to be mentored by Giulio Messori, the
CEO of Sweet Legal Tech, who provided invaluable guidance and insights and became a mentor in
my career journey.

5.5 My Role at Sweet Legal Tech

The LinkedIn advertisement depicted what Sweet Legal Tech expected from the role. The role I
applied to was listed as "Working Student - Copywriter and Social Media Planner (Italian
Speaker)". As the company navigated a phase of rapid expansion, there was a tangible need to
strengthen its social media presence and improve its copywriting strategy. The role was not just
about content creation but about understanding the essence of SLT, challenges, aspirations, and
pioneering stance in the legal tech discipline. The objective was clear: translating this intricate
understanding into compelling narratives that would cater to legal professionals and captivate a
broader audience, positioning Sweet Legal Tech as a beacon in the legal tech landscape.
At the time, Sweet Legal Tech's primary need was to increase brand awareness and strengthen its
brand identity. While the initial expectations were well-defined, my role at Sweet Legal Tech
quickly transcended that description. Even though most of my time was still spent crafting copies
and managing their social media, it became evident that the legal technology field was vast. Sweet
Legal Tech's ambitions were not confined to its growth but extended to the broader idea. I took over
the company's social media initiatives, crafting content that resonated with Sweet Legal Tech's
values and vision. However, as time passed and I delved deeper into the company, my
responsibilities expanded. I managed Sweet Legal Tech's communications and assisted other
startups in partnership with Sweet Legal Tech in articulating their vision and value proposition and
creating content whenever needed. This multidimensional role provided a holistic view of the legal
tech sector, its nuances, and its boundless potential.

81
5.6 Initial Challenges and Context
As mentioned before, upon joining Sweet Legal Tech as a 'Copywriter and Social Media Planner', it
became immediately evident that the scope of my role would be broader than the title suggested.
Despite its pioneering vision in the legal tech domain, Sweet Legal Tech faced challenges in its
communication strategy. While passionate about communication and social media, the CEO, Giulio,
primarily managed these channels without a formal marketing background, and a direct
consequence was the instability of the brand identity.

Brand identity is a set of tools, including core values, external expression, and marketing efforts
(Kapferer, 2012), that help a brand position itself in the mind of the targeted consumer (Aaker,
1997). As the company experienced exponential growth, the direct management of communications
from the CEO became increasingly untenable. This led to a dormant LinkedIn page that failed to
deliver any benefits for the company. An even bigger problem, which depends on the collaborative
approach to work, is that the brand identity needs to be appropriately portrayed on social media.
However, the content posted needed to be more cohesive. The brand identity of Sweet Legal Tech
needed to be portrayed adequately on its social media, and different tones of voice and color
palettes were used for each post, making it impossible for the consumer to establish a clear idea of
Sweet Legal Tech.

5.7 Key Responsibilities During My Internship

• Social Media Strategy: Revising Sweet Legal Tech's LinkedIn page was the most
challenging task. I began an extensive journey to learn and explore an efficient social media
strategy to achieve brand awareness. Initially, it was essential to know who we were trying to
target. This entailed determining the principal age and ethnic groups, educational attainment
levels, and interests of our target audience. I created some personas that would represent our
target and presented them to the team, as knowing our audience would enable us to adapt our
material better to appeal to them. Before incorporating new content, I examined previous
posts on Sweet Legal Tech's LinkedIn page. This research revealed what had previously
worked, what had not, and where there was room for improvement. With this basis, clear and
quantifiable targets were defined, such as growing page followers, enhancing interaction rates,

82
 or driving visitors to Sweet Legal Tech's website. These goals gave the plan direction and a
measure of its effectiveness. The development of content became essential to this strategy. I
planned the types of content to be produced, ranging from educational articles and case
studies to interactive polls and videos, guaranteeing a balance that would interest our audience
based on the insights from the audit and the set objectives. However, more than simply
posting content was needed. I planned ways to foster interaction on our posts by posing
questions, encouraging discussions, or collaborating with influencers in the legal tech space.
Encouraging feedback, input, and participation from the other members of the team and our
followers was pivotal. This direct line of communication helped refine our approach and
ensure that our content aligned with Sweet Legal Tech's brand values and objectives. This
meticulous approach entailed detailed planning, the creation of a rigid editorial calendar, and
the selection of graphics and templates that mirrored Sweet Legal Tech's style and vision. The
overarching goal was to increase visibility and position Sweet Legal Tech as a thought leader
and trusted resource in the legal tech domain.
• Content Creation: My role encompassed creating diverse content types, from whitepapers
and newsletters to blog posts and LinkedIn articles. These pieces aimed to illustrate Sweet
Legal Tech's efforts to keep up with the latest news and inventions in legal technology and the
broader technology landscape. Beyond copywriting, I researched and delved deep into the
sector's complexities, ensuring the content had a robust theoretical background.
• Graphic design: Besides creating the written content, I was responsible for the visual design
of posts, video content production, and any media shared on the LinkedIn page or the website.
This included covers for the newsletter, the articles, and the case studies presented on the
website.
• Market Research: A holistic understanding of the EU Legal Tech Market was crucial, as it
was to be continuously updated on the newest innovations. My research informed our social
media strategies, ensuring we catered to the niche audience while appealing to a broader
demographic.
• SEO Optimization: My main objective was to increase Sweet Legal Tech's online presence,
which prompted me to start an extensive SEO optimization approach. It began with thorough
keyword analysis, identifying phrases that prospective clients or legal technology aficionados

83
 would employ. I optimized our text using these keywords in mind, ensuring they were
naturally incorporated without compromising the quality of the writing. Titles and meta tag
descriptions were written with user-friendliness and search engine optimization in mind. I also
stressed the importance of internal linking, which directs users through a network of pertinent
content on our website and helps search engines grasp the content hierarchy. I ensured the
website was quick, responsive to mobile devices, and crawled error-free from a technical
standpoint. It was crucial to regularly evaluate results using tools like Google Analytics to
improve our strategy and respond to shifting SEO environments. In conclusion, SEO for
Sweet Legal Tech was an ongoing project that incorporated the complexities of search engine
algorithms with the specifics of the legal tech industry.
• Feedback and Continuous Improvement: Innovation and adaptability were at the heart of
my role. I constantly explored and tested new solutions and software to refine our content
strategy. Working for such an innovative company pushed me to search for better tools that
could improve my workflow continually. Regular feedback from the team was instrumental in
this iterative process.
• Performance Analysis: At Sweet Legal Tech, analyzing analytics was crucial to improving
our content strategy. I regularly checked the effectiveness of our postings using LinkedIn's
analytics capabilities. In order to determine the audience's effect and receptivity to our
material, we had to evaluate the reach and engagement by monitoring views, likes, shares, and
comments. I also looked at audience demographic breakdowns based on factors like area and
job function to better target our core audience. Additionally, I assessed the potency of various
material formats, ranging from articles to videos, enabling us to choose the most captivating
kinds. These analytics findings immediately drove our future material, ensuring that we
focused on the most popular themes and forms with our audience. By observing broader
engagement patterns over time, I helped Sweet Legal Tech adapt to shifting audience
preferences. This data-driven approach ensured our content strategy remained dynamic,
constantly adapting based on real-world feedback.
• Meta Ads: I employed Meta Ads for select campaigns to further amplify our reach. This
involved strategizing ad placements, budgeting, and analyzing ad performance to ensure
optimal ROI.

84
5.8 Tools Utilized
In today's digital age, the tools we employ play a crucial role in determining the quality and
efficiency of our work. During my internship at Sweet Legal Tech, I was introduced to a suite of
digital tools, each tailored to address specific challenges and tasks. I was always encouraged to
research new digital tools that I could utilize to facilitate and improve my job. These tools
streamlined my daily operations and significantly contributed to my professional growth. Here is a
closer look at each tool I have utilized and the skills I mastered through it:
• Notion: Notion is a comprehensive workspace platform that combines note-taking, task
management, and collaboration tools. Sweet Legal Tech served as the backbone for internal
communications, ensuring that the team remained aligned and informed, as only some team
members were always present in the office. Notion allowed us to maintain a well-structured
editorial calendar, ensuring timely content delivery. Beyond these conventional uses, I delved
into Notion's capabilities for website development. By leveraging its integrative features, we
could design a dynamic online website for Sweet Legal Tech, showcasing our offerings and
insights.
• Canva: Canva is a user-friendly graphic design platform offering many templates and design
elements. As my primary design tool, Canva enabled me to create visually striking posts,
covers, and other promotional materials. Navigating through its vast library of templates and
customizing them to resonate with Sweet Legal Tech's brand identity honed my design
sensibilities and ensured our digital assets were engaging and consistent.
• ChatGPT: ChatGPT, as profoundly explained in the literature review, is an advanced
conversational AI language model designed to provide detailed information and insights on a
wide range of topics. Recommended by my supervisor, it became an invaluable research
assistant, especially when I ventured into complex legal subjects. Utilizing ChatGPT, I could
craft well-informed posts, ensuring our content was accurate and high caliber. It also served as
a sounding board, helping refine my drafts for better clarity and impact.
• SEO Optimizers: Visibility is crucial in the enormous universe of digital. I used AI-driven
SEO optimization tools to ensure the material on Sweet Legal Tech got to the right people.
These tools gave me immediate feedback on the organization of the text, the use of keywords,

85
 and the meta descriptions, assisting me in making informed changes. Consequently, our
website continually received favorable rankings in search engine results, generating organic
traffic and improving our online presence.
• Super.com: Super.com is a tool that unifies the gap between content and web development.
Users may easily convert Notion pages into interactive web pages with this tool. My
knowledge of web building was broadened through my work with Super.com. I created
interactive web pages that clearly defined Sweet Legal Tech's goals and services by
combining Notion material with Super.com's online capabilities.
My skill set was strengthened by these technologies and the practical experience of using them in
real-life situations, which helped me to become competent at navigating the complex world of
digital communication.

5.9 Main Responsibilities During my Internship

Creation of the Brand Kit:
Creating a Brand Kit was fundamental to ensure the brand would be recognizable by consumers and
potential consumers who could encounter it online. Without a proper brand kit, companies risk
inconsistent branding across platforms, potentially leading to customer confusion, diminished brand
recognition, and a weakened brand image. Sweet Legal Tech already had a logo and a color palette,
fundamental tools for creating a brand identity (Aaker, 1997). However, they could have been used
more consistently and recognizably throughout their posts. Furthermore, there was not one physical
file collecting all these pieces of information, with the risk that whoever created the social media
content could not retrieve and utilize them properly. Therefore, one of my first tasks during my
internship was to create a proper brand kit and make it available to everyone in the company.

Redefining Brand Personality:

Legal technology, being a niche domain, posed unique challenges. The primary objective was to
engage those within the sector and to reach and educate a broader audience about the benefits of
integrating technology into legal practices. Sweet Legal Tech's vision was clear: they sought to
transition from a purely informative stance to an educative one on social media. The primary
objective of their new social media presence was not to push their consulting services but to create a

86
strong presence in the consumers' minds, positioning Sweet Legal Tech as a leading figure in the
innovation of the legal field. Furthermore, it was fundamental to create a distinctive brand
personality for Sweet Legal Tech that could highlight the main differences between all the
competitors: Sweet Legal Tech is a pioneer startup founded by a young and innovative lawyer under
40 years old, and the main goal it has always been to revolutionize the legal word from inside,
offering the unique perspective of professionals with both legal competences and technological
skills.

For brands to communicate their personality, values, and beliefs, it is essential to utilize the correct
tone of voice in copywriting, and it has even greater importance when talking about social media.
The brand's tone of voice, which expresses its personality and ideals, is essential in copywriting and
even more in defining the brand identity. (Aaker, 1997; Kapferer, 2012). It was a challenge to create
a unique tone of voice that could show Sweet Legal Tech's personality and part of my internship
was focused on this research. After some research, we could conclude that the target was the young
lawyers, law students of employers of legal department companies, who already were comfortable
with technology and could see the necessity of their workplace to improve their work-life balance
and speed up the bureaucracy around the legal processes. This required a complete overhaul of the
LinkedIn page, reshaping brand identity, and initiating a fresh, educational content strategy. For this
reason, we decided to use a more young and catchy tone of voice, utilizing memes and pop cultural
references.

Figure 2: LinkedIn Post 1

87
In this way, our primary goal was to differentiate Sweet Legal Tech from the main trend of legal
figures on LinkedIn, which usually post solemn and complex topics in a very descriptive way, and
target the younger generation by proposing the same topics in a different form, that might be more
appealing.

Content creation for the LinkedIn page:

During my six-month internship, I conceptualized and created several LinkedIn campaigns to
solidify Sweet Legal Tech's reputation in the legal domain. Here, I will present some of the most
successful campaigns:
• "Job Posts" Series: The 'job posts' series was among the most successful. We highlighted
three legal tech job opportunities in Europe every Monday, on-site or remote. This series
garnered significant attention, with over 30 episodes to its credit. It specifically targeted
people who wanted to start their career in the world of legal tech. However, it also spotlights
possibilities for professionals already in the legal world who might have desired to change
their role.
• Legal Tech Startup Insights: Another impactful campaign was "How legal tech businesses
were born". Each episode delved into the journey of a legal tech startup, occasionally
featuring interviews with its founders, providing valuable insights for those who were
thinking of starting a new business related to legal technology. In this campaign, my role was
to contact the startup marketing offices, proposing the collaboration and offering two possible
packages: one with some questions for the CEO and another with a more general description.
Whenever they chose the interview, my task was to transform the answers of the CEO into a
clear, catchy, and explicative carousel post (Figure 3). If the CEO could not give us the
interview, but the company still wanted to collaborate, the marketing office would send us a
marketing pack, where I could find information about the company's foundation and create
the post. (Figure 4)

88
 Figure 3: LinkedIn Post 2

Figure 4: LinkedIn Post 3

• Legal Tech Courses Overview: The 'legal tech courses' campaign was pivotal in our
educational outreach. It cataloged critical legal tech courses globally, with each episode
offering a detailed look into the course's content. Part of my task was to research these
courses, insert them in the Index, and subsequently create the post. This campaign targeted
law students who might have wanted to integrate a Legal Tech Course into their academic
journey.
In addition to the structured campaigns I managed, I was also responsible for generating a variety of
content on a near-weekly basis. This ranged from updates on the latest developments in the legal
technology world to informative articles that provided insights into the industry. Occasionally, I
crafted lighter, humorous posts to engage our audience differently.
Though unique in its message and length, each piece of content had to adhere to a consistent visual
theme. This was crucial for establishing and maintaining a strong brand identity on social media
platforms. Despite the inherent differences between individual posts or series, we needed to present
our followers with a cohesive and instantly recognizable visual style. This consistency enhanced our

89
brand's professionalism and ensured that our content was immediately identifiable amidst the
plethora of information on social media.
The newsletter:
Newsletters are a potent tool for organizations to communicate directly with their audience, offering
insights and fostering a deeper connection with previous followers while reaching new potential
followers. The "2030 Lawyers" newsletter, initiated by Sweet Legal Tech, was designed with a
visionary objective: to equip legal professionals for the transformative shifts in the legal landscape.
While this newsletter's concept was established before my tenure, I had the opportunity to craft the
content for the episodes launched during my internship. I began curating five "2030 Lawyers"
episodes from October to March. Each edition delved into pertinent topics in the world of legal
technology. Moreover, the newsletters often spotlighted case studies, illustrating the tangible
benefits of integrating legal technology software into conventional workflows. These real-world
examples served as compelling narratives, demonstrating the tangible efficiencies and
advancements that technology brought to the legal sector.The newsletter's long-form content format
was particularly advantageous for educational purposes. It allowed for a comprehensive exploration
of topics, ensuring readers received a well-rounded understanding. Furthermore, by sharing these
newsletters on LinkedIn, we expanded our reach, tapping into a broader audience keen on
understanding the confluence of law and technology. The initiative of utilizing LinkedIn as a tool
for promoting the newsletter led to more than seven hundred subscribers on the platform. This
strategy amplified our brand presence and positioned Sweet Legal Tech as a thought leader
committed to enlightening the legal community about the future of their profession.

Revitalizing Sweet Legal Tech's Online Presence: The Website

The Sweet Legal Tech blog featured the articles concurrently with creating and publishing weekly
material, ensuring that our insights reached a broad audience across several platforms. The website,
which served as the digital representation of the organization, was crucial in forming opinions and
giving visitors access to helpful information. As part of my duties, I had to update the blog and
media coverage area consistently.

90
The requirement for a more thorough website overhaul became clear as my internship continued.
The choice to partially remodel the website provided me with a rare chance to explore previously
uncharted territory in my professional development. I actively participated in the copywriting
process to make sure the material reflected the brand personality and mission of Sweet Legal Tech.
In order to create a user-friendly and aesthetically pleasing digital area, we used technologies like
Super.com and Notion. These platforms offered a combination of usefulness and aesthetics.
As my internship ended, I started working on another vital project: website SEO improvement. A
website's visibility must be improved through search engine optimization in order for it to be seen
by the target audience. During my internship, I actively participated in team events at Sweet Legal
Tech in addition to my core responsibilities as a copywriter and social media manager. I interacted
with many team members and gained a more comprehensive view of how the business operated as a
result. We held "Stand-up" meetings each week. In these sessions, everyone discussed their current
projects, weekly goals, and any assistance they might need. It was a fantastic team communication
tool to motivate one another. Additionally, my supervisor and I had weekly one-on-one meetings.
We thoroughly addressed our social media plans throughout these meetings. We evaluated what was
effective and needed improvement before deciding on the week's tasks. These sessions ensured I
was on the correct track and helped me better grasp our objectives. I also got to attend meetings
with law firms and partner companies. This experience was beneficial since it broadened my
perspective of the legal technology sector. A pleasure was getting to see how the PR firm handled
CRCLEX's social media. Observing their strategy gave me fresh suggestions and thoughts for our
tactics.

91
5.10 Impact of the LinkedIn Campaigns
When I first joined Sweet Legal Tech, the company's LinkedIn page existed, but its potential still
needed to be fully discovered. Throughout my six-month tenure, I contributed to a significant
transformation in the page's reach and engagement. My content strategy was diverse and tailored to
the platform's audience; it curated a mix of articles, visually appealing graphic posts, interactive
polls, and informative videos. This multifaceted approach ensured that we catered to the varied
interests of our followers and kept them consistently engaged. The results of these efforts were
tangible. In just half a year, the LinkedIn page saw an influx of over 500 new followers. More
impressively, the engagement metrics showcased the effectiveness of our strategies. The average
number of organics views per post went from 100 to 500. Similarly, the average reactions per post
experienced a significant surge, moving from a mere one or two to approximately 15. This
heightened engagement indicated increased visibility and a deeper connection with our audience.
Furthermore, our newsletter on LinkedIn also garnered substantial attention. By the end of my
tenure, we had amassed a subscriber base of 700 individuals, a clear indication of our content's
value to professionals in the field. However, while the LinkedIn campaign's impact was evident, the
influence of our efforts on the website remains challenging to define appropriately. The website's
optimization and content revamp are still ongoing projects, so a comprehensive assessment of their
impact will only be possible once they are fully realized. Nonetheless, the strides made on the
LinkedIn front provide a promising outlook for the website's future performance.

5.11 Transition and Future Endeavors

My internship came to a successful and optimistic conclusion. Sweet Legal Tech offered me a
freelance collaboration and proposed the Head of Communication and Social Media Manager role
after recognizing my work and the value I provided to the company. Furthermore, I started a
freelance collaboration as a Brand Identity specialist and social media manager for other
organizations in partnership with SLT, such asFamTech, and with the law firm CRCLEX, utilizing
my knowledge to help them develop their communication strategy and strengthen their market
presence. This served as evidence of the abilities I had created and the influence I had produced.

92
5.12 Critical Reflection on My Internship Experience
A startup internship offers a unique combination of possibilities and difficulties. I was given an
unmatched introduction to the complex world of brand communication at Sweet Legal Tech. In
contrast to the segmented jobs sometimes seen at more prominent businesses, a startup environment
enabled me to fill many responsibilities while fluidly transferring from one duty to another,
preventing me from feeling constrained in a minor position. In this dynamic environment, I could
put my academic competencies into practice and see how my efforts had a real-world impact.
One of the most enriching aspects of my internship was the privilege to work closely with the
company's founder and a team of seasoned professionals from diverse domains. This proximity to
the decision-making nucleus of the company provided me with invaluable insights and fostered my
personal growth. Furthermore, it allowed me to understand the company's needs and utilize my
skills to help them. Overall, I have positively impacted the creation of brand awareness for Sweet
Legal Tech on social media.

However, every silver lining has its own cloud. Since startups are lean teams with broad
responsibilities, the training I got was more practical than academic or connected to my educational
background and I still didn’t have full-time marketing related mentoring. My supervisors mainly
had experience in fields unrelated to marketing. This occasionally made getting expert advice and
comments difficult, and it had to offer the educational experience required for an internship.
The startup's variable role structure also provided its share of difficulties. Although it gave me a
wide range of skills, the absence of a clearly defined function occasionally left room for confusion
over my duties. A more organized training program that concentrated on particular activities may
have added greater clarity. I ran into a constraint where I had to provide legal-related content. I
frequently had to rely on others to give me insight into the material because these subjects, while
quite accessible to legal professionals, were difficult for someone without a foundational
background in law. Even when I could work on them independently, navigating these topics
required extensive research and a keen understanding to ensure accuracy and relevance.

Looking back, I can see how valuable my internship at Sweet Legal Tech was. While it may not
have been a typical educational experience rooted in theory, it thoroughly explored the operational

93
facets of brand creation and social media management. My professional perspective has been
irrevocably changed by the possibilities and difficulties I have faced, and I now have the knowledge
and skills I need to navigate the ever-changing world of digital communication successfully.

94
6. Methodology

This chapter thoroughly explains the methodology employed to scrutinize the effects of artificial
intelligence on copywriting and people's ability to discern between human and AI-generated
content.The chapter introduces the research questions guiding this investigation before explaining
the research design. In this study, we chose a mixed technique that combines qualitative and
quantitative methodologies because it allows a thorough investigation of the problem. Additionally,
the decision to use an experiment and a focus group is explained, highlighting the justification for
this decision. The chapter concludes with a thorough study of the participant's demographics,
providing a solid framework for the subsequent research.

6.1 Research questions

A growing body of research explores the impact of artificial intelligence on various types of writing,
from creative writing (Dargham et al., 2022; Ippolito et al., 2022; Klein et al., 1973; Turner, 1993)
to academic writing (Brown et al., 2022; Shahriar & Hayawi, 2023; Mosca & Such, 2022; Silva et
al., 2020), there is a notable gap in the literature explicitly focusing on the domain of copywriting.
Most existing studies have centered on the broader implications of A.I. in content creation
(Dargham et al., 2022; Ippolito et al., 2022), language translation, grammar checking (Clark et al.,
2018), and creative writing (Dargham et al., 2022) but have not delved deeply into the nuances and
specific challenges associated with copywriting.

Therefore, to address this gap, this study aims to delve into the transformative impact of artificial
intelligence on copywriting. The researcher's background in copywriting and experience with
artificial intelligence gained during the internship started an investigation into the capabilities of
advanced A.I.-powered language models, such as GPT-3.5, and its application in ChatGPT as of
November 30, 2022 (OpenAi, n.d.). With their incredible skills, these large language models are
confusing the lines between human and A.I.-generated content, raising questions about trust,
authenticity, and ethics in copywriting.

95
The primary research question is: What is the prospective influence of A.I. on the sphere of
copywriting? This question seeks to understand A.I.'s potential role in reshaping content creation,
the future of the copywriting profession in an A.I.-centric world, and the ethical and societal
implications of A.I.-generated content.
To delve deeper and provide tangible insights, a secondary investigative question arises: Can
people discern between human-crafted and A.I.-generated content? This question is pivotal in
understanding the real-world implications of A.I. in copywriting. If users cannot differentiate
between human and A.I.-produced content, the ramifications for the copywriting profession could
be significant.

Two key hypotheses drive this research:

1. Suppose users cannot distinguish between human-authored and AI-generated content on
social media. In that case, the influence of Large Language Models, like GPT-3.5, will be
profound, potentially jeopardizing the traditional role of copywriters.
2. If consumers recognize AI-generated content, the impact will still be present, but it will be
minor. and it will depend also on the emotional and perceptual response of the consumers
Do they feel deceived, indifferent, or intrigued?

Furthermore, the study explores the implications of AI-driven copywriting on social media
marketing, focusing on platforms like LinkedIn and Instagram. By addressing these research
questions and hypotheses, the study aims to contribute significantly to the discourse on A.I. and
copywriting, shedding light on the potential benefits, challenges, and societal impacts of A.I.'s
growing presence in the industry.

6.2 Research Design

A mixed methods approach was chosen for this study, aligning with Creswell's definition of mixed
methods research as a type of research that collects, analyzes, and combines both qualitative and
quantitative research methods in a single research framework to understand a research problem
(Creswell et al., 2003; Creswell, 2014). This approach has its roots in the historical transitions in
research methodologies, with its inception in psychology and the multitrait-multimethod matrix.

96
Over time, mixed methods research has garnered significant interest across various disciplines, with
numerous publications and funded projects dedicated to its exploration (Creswell et al., 2003).

The decision to employ a mixed-methods approach was multifaceted. Primarily, it allowed for a
comprehensive understanding of the research questions, drawing from the strengths of both
quantitative and qualitative paradigms. This dual approach facilitated a deeper understanding of
individual experiences and enabled the generalization of findings to a broader population,
addressing the research topic holistically. Furthermore, the mixed-methods approach ensured the
validation and verification of findings, producing well-validated knowledge claims backed by
robust evidence. The pragmatic nature of mixed methods allowed for diverse data collection and
analysis strategies, effectively addressing various research nuances (Creswell et al., 2003).

The qualitative component of this research involved focus group discussions, offering a platform to
delve deeper into participants' perceptions regarding AI-generated content. This component was
executed through two focus groups, ensuring a balanced representation of millennials and
Generation Z individuals. The discussions were structured, beginning with an ice-breaking phase
and culminating in an in-depth conversation about participants' experiences with AI-generated
content. Simultaneously, a quantitative experiment was conducted, employing a comparative design
to measure participants' accuracy in distinguishing between human and AI-generated content. This
component complemented the qualitative insights, providing a broader perspective on participants'
abilities.

To enhance the study's credibility, several measures were implemented. The AI-generated content
was crafted using the GPT-3.5 language model, ensuring quality and coherence. Both AI and
human-generated content were presented as anonymous posts to eliminate bias, recreating a
LinkedIn or an Instagram post. References to brands or names that could be recognized have been
hidden.

In conclusion, this mixed methods study comprehensively explored A.I.'s impact on copywriting.
Integrating qualitative and quantitative methodologies achieved a richer understanding of the

97
research topic, ensuring data validity and reflecting participants' accurate perceptions of AI-
generated content.

6.3 Triangulation in Research

Triangulation, as adapted in research, signifies multiple approaches to investigate a research
question and enhances confidence in findings by corroborating propositions through two or more
independent measures (Heale & Forbes, 2013). This approach offers a more holistic view of the
results, ensuring a comprehensive understanding of the research topic (Tashakkori & Creswell,
2007). Historically introduced into qualitative research in the 1950s, triangulation was conceived as
a strategy to mitigate potential biases arising from the sole use of one method (Heale & Forbes,
2013). Its application can span multiple theories, data sources, methods, or investigators within a
single study, allowing versatility (Heale & Forbes, 2013; Tashakkori & Creswell, 2007).

The most prevalent form of triangulation is methodological triangulation, which may encompass
multiple sets of data collection using the same methodology or different methodologies, such as
qualitative and quantitative (Heale & Forbes, 2013). Triangulation, mainly when used in mixed
methods research, can yield three potential outcomes: converging results that validate the findings,
complementary results that highlight different facets of the phenomenon, or divergent results that
can pave the way for new, enhanced explanations. For instance, a study might employ a survey to
gauge service user satisfaction and focus groups to delve deeper into their perceptions, linking and
comparing these methods to address a unified research question (Heale & Forbes, 2013).

However, the process of triangulation has its challenges. While it is lauded for adding depth and
richness to research inquiries, it operates on the assumption that data from distinct research methods
are comparable, which may only sometimes be the case (O'Cathain et al., 2010). Moreover,
convergent findings from multiple data sets might need to be more accurate if each set is inherently
flawed. Despite these potential pitfalls, the overarching consensus in the research community is that
triangulation, when properly executed, can significantly enhance the depth of a study, providing a
more nuanced understanding of the phenomenon under investigation (Heale & Forbes, 2013;
O'Cathain et al., 2010).

98
In pursuing a comprehensive understanding of the research topic, this study employed
methodological triangulation, integrating qualitative and quantitative research methods. The
qualitative component, consisting of focus group discussions, and the quantitative experiment
assessed participants' ability to differentiate between AI-generated and human-written content. The
data were gathered concurrently. This simultaneous data collection ensured that insights from one
method could be immediately juxtaposed with the other, providing a dynamic and multifaceted
view of the research question.

While both methods offered invaluable insights, the data from the focus group discussions held a
distinctive edge in depth and nuance. The discussions emphasized the participants' abilities to
discern content origins and delved into the underlying biases, perceptions, and emotions that
influenced their judgments. The richness of the qualitative data, characterized by participants'
candid reflections and interactions, provided a deeper understanding of their experiences and
perspectives. This depth was especially pivotal in discerning the subtleties of participants' biases,
their preconceived notions about AI-generated content, and the emotional undertones that might
have influenced their content evaluations.

On the other hand, the quantitative experiment offered a broader, more generalized view of
participants' discernment abilities. However, with the context and depth provided by the focus group
discussions, these quantitative findings needed more nuance and depth to understand the
phenomenon entirely. While the quantitative data provided the "what" regarding participants'
abilities, the qualitative data from the focus groups offered the "why" – the deeper motivations,
biases, and feelings underpinning their responses. Finally, the combination of these two methods,
with particular attention to the insights from the focus groups, ensured a complete understanding of
the research topic, highlighting the intricate interplay between participants' abilities, biases, and
emotions in the realm of AI-generated content.

99
6.4 Target
This research primarily focuses on individuals actively engaged on social media platforms and
frequently encountering copywriting content through online interactions with brands. Given the
increasing integration of artificial intelligence in copywriting, gathering insights from a diverse
demographic is imperative to comprehensively understand its impact on content creation and the
perceptions surrounding it. This study aims to discern if there are variations in these perceptions
based on factors such as occupation and educational background. Recognizing the significance of
AI-driven technologies and their potential to revolutionize the copywriting domain, this research
emphasizes understanding the perspectives and tendencies of the younger groups, specifically
millennials and Generation Z.

There are compelling reasons to focus this study on the younger generations that are already active
on social media and interacting with brands, which includes both millennials and Generation Z.
These generations are very engaged across a variety of digital channels, making them the critical
consumers and interactors with online material. Younger generations could be already active on
social media, but they are too young to have a clear vision of A.I. or a rational vision of a brand's
communication, and the goal of studying these groups is to acquire insight into their content
choices, which are frequently impacted by more significant socio-cultural issues. Notably,
"Generation Z will represent 40% of the market's consumers in 2020" (Serbanescu, 2022, p.68).
This underscores the urgency to understand how to target them effectively and what strategies
companies should utilize to capture their attention. Given their substantial market presence, it
becomes imperative to discern if A.I. can genuinely resonate with them or if they can see through
AI-generated content and which reaction they could have.

Furthermore, advances in artificial intelligence and language modeling are expected to have the
most significant influence on millennials and Generation Z. Serbanescu (2022) emphasizes that Gen
Z, often known as "digital natives," are those who "fluently speak the language of computers, the
Internet, phones, and video games." Unlike millennials, they have grown up with cell phones and
are used to handling many gadgets. This familiarity and proficiency with technology emphasize the
need to know their involvement with and impressions of AI-generated copywriting. Their

100
perspectives can illuminate possible issues, possibilities, and ethical concerns linked with AI-
enhanced copywriting approaches. Furthermore, millennials and Generation Z have distinct forms
of digital intelligence. They grew up in a sophisticated digital world that forced them to adapt and
succeed. This "digital intelligence," is the result of their constant engagement with digital media and
is promoted by a style of learning known as "learning by doing" (Serbanescu, 2022, p. 63). This
intellect and active participation in digital realms make them ideal candidates for our research.

It is critical to underline the significance of incorporating people of all demographics in our study.
Collecting various perspectives can strengthen the findings by guaranteeing a varied sample. As the
primary purpose of this research is to investigate the capacity of social media users to notice A.I.-
powered copywriting, it is essential to understand if an individual's academic background or
professional position affects their ability to recognize AI-generated material. The target sample must
be diverse, including people from many academic areas and professional sectors, to gain a
comprehensive understanding. This method ensures that our findings are representative and can
provide insight into how different parts of the population engage with and interpret AI-driven
material.

6.5 Sample of Participants

In social science research, studying entire populations is rarely a possibility. Instead, a subset or
sample is typically examined to represent the broader population. This sampling approach is
adopted when it is impractical to involve the entire population or when identifying all target group
members is challenging (Taherdoost, 2016).

For this study, participants were chosen using the convenience sampling method for the experiment
and focus group discussions. Convenience sampling is non-probability sampling (Vehovar et al.,
2016). Unlike its probability counterparts, this method is less objective, as it does not give every
individual in the target population an equal chance of participation. Other standard non-probability
sampling techniques include purposeful sampling, snowball recruiting, and quota sampling, a
variant combining convenience sampling with systematic population segmentation (Vehovar et al.,

101
2016). It is essential to note that findings from non-probability sampling cannot be generalized to
broader populations. Any causal relationships or empirical associations identified are only relevant
to the study's participants. Such sampling methods also do not allow calculating sampling errors or
data accuracy concerning the entire population. Despite these limitations, convenience sampling
remains popular in clinical and qualitative research due to its cost-effectiveness, efficiency, and
simplicity (Stratton, 2021). The decision to employ convenience sampling was driven by practical
constraints, including time, resources, and the necessity to engage participants who were readily
available within the stipulated period. Convenience sampling might result in non-representative
data, reducing the generalizability and validity of the study. The lack of random sampling may
induce bias and restrict the findings' application in other circumstances. The results of this method
should be interpreted with caution, particularly when establishing causal conclusions (Simkus,
2022; Stratton, 2021).

However, efforts were made to include a diverse sample regarding age, educational attainment, and
other pertinent variables to ensure the results were applied to a larger population. The research
offers a complete picture of the attitudes and behaviors of the broader community by involving
various people from various backgrounds. This strategy is beneficial when attempting to derive
findings that may be used in a broader context. Furthermore, using a varied sample might show
patterns and insights not seen in a more homogeneous group, resulting in a more nuanced
knowledge of the subject matter under inquiry.

To select the participants, the eligibility criteria included individuals aged 18 and above who were
fluent in English and active on social media platforms. The selection process for the focus groups
was carried out systematically to ensure the participants represented a cross-section of the
population. Additionally, participants were selected from different regions across Europe to ensure a
more comprehensive representation. Meanwhile, it was more challenging to control the participants
of the survey. However, efforts were made to ensure the spread among different European countries
and different works and academic environments. The selection of participants fluent in English was
essential to ensure that the data collected was high quality and could be easily understood. Although
participants in this study had to be fluent in English, there might be minimal potential bias due to

102
language proficiency. The moderator provided clear instructions and explanations to ensure that
participants understood the study's questions and tasks. In the focus group, a translator was also
available for those who needed help understanding and responding to materials in their native
language. These measures aimed to promote inclusivity and enhance the study's validity and
comprehensiveness. Overall, the selection process was carried out with maximum care to ensure the
study's findings were robust and representative of the wider population.

The focus group sample for this study consisted of 12 participants, with ages ranging from 22 to 34
years, representing a diverse international spectrum hailing from seven different countries. The
online experiment collected 514 responses. All the participants were Millenials Or Gen Z and from
European countries. The respondent's level of education was between having a high school diploma
or equivalent to having completed a doctorate or Ph.D. All the participants were active on social
media. This demographic distribution ensures a broad cultural perspective and a rich mix of views
from millennials and Generation Z participants, enhancing the richness and relevance of the study's
findings.

6.6 Texts Used

Participants were presented with five pairs of texts in the focus groups and the experimental
examination, each labeled 'A' or 'B'. Within each pair, one text was an original LinkedIn or
Instagram post from a legal tech-related company, while the other was an AI-generated version of
the same post, produced using the GPT-3.5 language model. To minimize potential bias, brand
names were obscured. However, visual cues indicative of the original social media platforms were
maintained, enabling participants to determine the intended target audience of the post. In the
experimental study, participants were tasked with distinguishing between human-written and AI-
generated texts by assigning the appropriate label to each text. During the focus group sessions,
participants were asked to identify the origin of each text and then prompted to critically analyze
various aspects of the samples, such as tone, style, grammar, and vocabulary.

103
Set 1:

Figure 5: Set 1, Text A

Human Generated

Figure 6: Set 1, Text B

A.I. Generated

104
Set 2:

Figure 7: Set 2, Text A

A.I. Generated

Figure 8: Set 2, Text B

Human Generated

105
Set 3:

Figure 9: Set 3, Text A

Human Generated

Figure 10: Set 3, Text B

A.I. Generated

106
Set 4:

Figure 11: Set 4, Text A

Human Generated

Figure 12: Set 4, Text B

A.I. Generated

107
Set 5:

Figure 13: Set 5, Text A

A.I. Generated

Figure 14: Set 5, Text B

Human Generated

108
7. Findings

This study delves into the transformative potential of artificial intelligence within copywriting. With
the advent of AI-powered language models like GPT-3.5, exemplified in ChatGPT, a potential shift
in copywriting is anticipated. Guided by the primary question, "What is the prospective influence of
A.I. on copywriting?" and the secondary question, "Can people discern between human-crafted and
AI-generated content?" this research aims to augment the discourse on A.I.'s role in copywriting,
highlighting its merits, challenges, and societal implications.

7.1 Focus groups

Two focus groups were set up to investigate how people perceive and distinguish between
copywriting for social media created by humans and that generated by artificial intelligence. During
these sessions, participants were presented with five pairs of texts; each marked as 'A' or 'B'.
For each pair, one of the texts was a real copy previously posted on social media by a company,
while the other was an AI-recreated version of the exact text. All AI versions were produced using
the same language model, GPT-3.
All brand names included in the texts and the company that originally posted the copy were hidden
to avoid potential bias. The participants were asked to identify which texts were written by humans
and which were generated by AI. They were encouraged to examine and discuss various elements of
the samples, such as tone, style, grammar, and vocabulary.

7.2 Focus Group 1

The first focus group was composed of six participants, each bringing unique perspectives to the
conversation. The group was a combination of international professionals from various sectors and
backgrounds. Firstly, Participant 1, a 26-year-old student studying communication, marketing, and
advertising. Participant 2, a 30-year-old customer behavior marketing specialist, offered insights
from his professional experience in understanding consumer behavior. Participant 3, the youngest
participant at 23, contributed her experience from the human resources sector, offering a different
angle on the topic. Participant 4, a 26-year-old software engineer, introduced a technical standpoint
into the dialogue thanks to his software and AI development background. Participant 5, a 27-year-

109
old marketing professional specializing in partnerships, added another layer of depth to the
marketing perspectives within the group. Lastly, Participant 6, a psychology student, provided
invaluable insights into human behavior and the psychological aspects of communication. This
group, ranging from their early 20s to early 30s and hailing from various European countries, was
immersed in roles that frequently interacted with text composition and communication in both
academic and corporate settings. Their diverse backgrounds and areas of expertise facilitated a rich,
enlightening conversation about the role of AI in copywriting and its capability to generate social
media text.

The focus group began with a structured introduction where the facilitator set the rules. Participants
were advised to avoid speaking over one another and to listen actively to their peers. The facilitator
then provided a comprehensive overview of the topic: artificial intelligence (AI) and its application
in copywriting, particularly in social media post generation. The participants were briefed on the
experiment's structure: they were asked to identify which texts were written by humans and which
were generated by AI. They were encouraged to examine and discuss various elements of the
samples, such as tone, style, grammar, and vocabulary.

This part of the discussion aimed to explore the participants' preconceived notions about AI and its
potential impact on the world of copywriting. Participant 1 expressed apprehension about AI,
suggesting it might lead to job losses in the copywriting domain. Participants 3 and 4, on the other
hand, viewed AI as a complementary tool, emphasizing that copywriters should adapt and learn to
harness AI's capabilities effectively. This sentiment of AI as a supportive tool rather than a
replacement was further echoed by Participant 5, who believed that while AI could assist in specific
tasks, the human touch in copywriting remained irreplaceable.

The discussion then transitioned to participants' personal interaction with social media. They were
asked about their preferred type of content (written or visual), their following of brands or
companies on social media, and their ability to recognize a brand's distinctive tone of voice. This
segment revealed a diverse range of preferences and habits among the participants. While most
leaned toward visual content, others had a predilection for written material; in particular,

110
Partecipant 1 and Partecipant 3 expressed their preferences for Twitter's content. All participants
concurred that they followed brands or companies on social media and felt confident in their ability
to discern a brand's unique tone of voice.

Subsequently, the conversation pivoted to participants' experiences and thoughts on AI-generated

content they might have encountered, especially from brands they follow on social media.
Participant 3 noted that while he could not definitively identify AI-generated content, he believed he
had encountered it multiple times, associating its well-crafted nature as a potential indicator.
Participant 6 offered a more nuanced perspective, suggesting that the content's distinctiveness and
quality played pivotal roles in determining its origin, whether AI or human. The moderator then
steered the conversation towards the intricacy of language used in the texts. Participants delved into
a detailed discussion about the complexity of words and the ease of reading as potential indicators
of AI authorship.

Subsequently, the focus group proceeded with the second part: the analysis. After reading one set of
texts at a time, and once they formulated their initial impressions, they were asked to determine
which text was crafted by a human author and which was generated by AI.
• Set 1: The primary discussion revolved around the believability of the first text being written by a
human. Many felt its simplicity and straightforwardness made it seem human-written. The second
text, with its detailed explanation and complex word choice, was primarily believed to be AI-
generated. It was perceived as more formal and distant, contrasting with the personal touch of the
first text. The revelation that the second text was human-authored took the group by surprise.
• Set 2: The second set saw the participants divided in their opinions. They found it challenging to
discern between the two texts, noting the complexity of words in the first and the simplicity of the
second. Some felt the second text's structure, especially its explicit call to action, made it more
user-friendly. Therefore, they believed a human copywriter was more likely to write it. The layout
of the text became a discussion point, with varying opinions on AI's ability to produce uniform
text. Certain words, like "swiftly," were highlighted, with one participant noting, "It's not a
common word to use. It sounds professional, but as someone who isn't a native English speaker,
some terms feel limited to me. That's the impression I get from this text." The overarching

111
 sentiment was that the first text's professional jargon made it seem AI-generated, though the
context of the target audience being lawyers on LinkedIn was considered.
• Set 3: For the third set, the group was evenly split in their judgments. A notable point from the
discussion was the association of low-quality content with AI. Some participants felt that even if
targeting professionals like lawyers, content should be more concise and user-friendly to engage
the reader effectively.
• Set 4: The majority correctly identified the AI-generated content in the fourth set. The discussion
highlighted two main points: the unique choice of emojis by the AI, which were not commonly
used, and the captivating nature of the text. Both these factors contributed to the perception of AI-
generated content being higher quality, and this last point generated a conflict with what the same
participants believed before.
• Set 5: The final set saw no participants correctly identifying the AI-generated content. The
human-authored text's formal and structured nature deceived them. A significant discussion point
was the importance of a consistent tone of voice across posts. The participants felt that detecting
AI authorship became challenging if a post maintained the same style as previous ones.

Participants compared notes on the characteristics they associated with human and AI-generated
copy, highlighting the nuances they observed and the assumptions they made. Undoubtedly, the
participants approached this task from different angles, reflecting their diverse backgrounds and
perspectives. In discerning between human and AI-generated text, participants grappled with a
complex array of factors. The discussion unfolded in many layers, uncovering how participants
engaged with and understood the texts. One of the predominant themes from the discussion was the
element of emotionality and depth of expression in the texts. Several participants posited that the
capacity to convey genuine sentiment and depth remained a uniquely human trait. They perceived
an elusive human 'spark' in the texts they believed were human-written, citing this as a key
differential between human and AI writing. Conversely, other participants associated mechanical
precision and correctness in language use with AI. They pointed out that while an AI could
construct grammatically perfect sentences and use buzzwords correctly, it might fail to successfully
replicate the quirks and idiosyncrasies that often characterize human writing. Furthermore,
structural complexity and logical flow of narratives also emerged as significant factors. Participants

112
argued that despite its computational power, AI might still need help to weave a coherent and
cohesive narrative from start to finish. Disjointed structures or random ideas could be signs of an AI
attempting to mimic human thought processes. Concluding the discussion, the group acknowledged
the blurred boundaries between human and AI writing. The viewpoints emphasized the participants'
different assumptions about AI capabilities, shedding light on the intricate dynamics between
technology, language, and human perception. Given the rapid advancement of AI, identifying AI-
generated content solely based on these factors might become increasingly challenging. In the
revelation phase, the results of the exercise were shared.

This prompted another layer of conversation as participants reflected on their accuracy in

identifying the origins of the texts. Some expressed surprise at the results, while others discussed
the factors that may have led them to correctly or incorrectly guess the texts' origins. The focus
group concluded with a post-results discussion where participants discussed their perceptions of AI
in copywriting in light of the exercise. They considered signs indicating a text was AI-generated,
debated how their views on AI in copywriting may have changed, and speculated how they might
approach similar texts in the future. General thoughts on AI and its implications were also shared,
with participants recognizing AI's increasingly prominent role in various fields, including
copywriting. The discussion underscored the importance of understanding AI's capabilities and
limitations in generating human-like text.

7.4 Focus Group 2

The second focus group was a diverse mix of professionals from various fields and backgrounds.
Participant 7, a 25-year-old graphic designer from Milan, brought a creative perspective. Participant
8, a 27-year-old AI specialist from Spain, provided technical insights. Participant 9, a project
manager from Vienna, shared project management and organizational viewpoints. Participant 10, a
28-year-old marketing specialist from Germany, contributed marketing expertise. Participant 11, a
34-year-old masseur from Portugal, offered a unique perspective from the wellness industry. Lastly,
Participant 12, a 25-year-old account strategist from Italy, enriched the discussion with strategic
insights.

113
The discussion delved into the attendees' AI assumptions. While some saw artificial intelligence as
a democratizing force in content production, others were concerned about its potential to supplant
human roles. One attendee emphasized the value of AI tools such as Grammarly, particularly for
non-native English speakers and those with dyslexia. They explained how such tools may help with
text organization, creativity, and grammatical precision. The focus, however, was on the
indispensable human aspect of copywriting, namely the empathy and emotions that individuals put
into their work. The group also considered the possibility of AI making linguistic errors. Some feel
that AI is more prone to errors, while others believe humans are more prone to errors.
The group's social media habits were explored in-depth. All participants were active on social
media, with most preferring visual content. However, some expressed a preference for written
content. When asked about their interactions with brands on social media, the group unanimously
agreed that they followed various brands. However, the group did not unanimously believe they
could recognize the unique tone of voice of the brands they followed. One participant shared her
experience using AI to craft catchy captions for a company's social media posts. Interestingly, she
noted that no one ever discerned a difference between her AI-generated captions and those she
wrote herself, and these confirmed the opinion of the participant that it would be hard to notice a
shift in the tone of voice of a brand if it is not very particular. The participants discussed their
potential encounters with AI-generated content on social media. While some believed they had
encountered such content due to its catchy nature, others needed clarification. The group
acknowledged the increasing prevalence of AI in content creation, emphasizing its potential to
produce high-quality content. However, they also highlighted the importance of the human touch in
crafting engaging and emotionally resonant content.

The discussion also touched upon the perception of quality associated with AI-generated content.
One participant remarked, "I don't think like AI means good quality. For me, at least. It can work
also in the opposite. If you see something written really bad, do you think it's a bad copywriter or a
bad user?" This sparked a debate within the group, revealing divided opinions. While some believed
that AI could produce high-quality content, others felt that poor quality could be attributed to either
a subpar AI model or a human copywriter lacking proficiency.

114
"I've seen some crazy things." "Bad things or good things?" "Bad things. But with chatGPT, I mean,
I'm impressed by it because it has so much information. Like sometimes I'm just like, wow, crazy
that it knows everything. But the way it's written, I don't know."
The group acknowledged the variability in AI's output quality, emphasizing that it sometimes only
guarantees excellence.

The second focus group delved into analyzing the provided text sets, aiming to discern between
human and AI-generated content. Their discussions and conclusions were as follows:
• Set 1: The unanimous consensus was that a human wrote the first text, and they were correct. The
primary reason cited was the text's concise nature and limited use of emojis. The group felt that
the AI text's use of emojis was excessive and "fake" and did not resonate like a human would
typically do.
• Set 2: Opinions were split evenly among the participants. Some believed the first text was
human-generated due to its lengthy and slightly convoluted structure, which they felt compared to
a human's train of thought. They also felt it had a more persuasive tone, similar to the usual
advertising language. Conversely, others felt the second text was more refined and user-friendly,
pointing out its clear structure and call to action and identifying it as a sign of a human hand.
Some saw the dense paragraph format of the first text as a sign of machine generation, as humans
tend to break up their writing for better readability.
• Set 3: All participants mistakenly believed the first text was AI-generated. The placement of
hashtags within the text was a significant point of discussion, with the group agreeing that
humans are more likely to incorporate hashtags seamlessly into their writing. Some participants
also felt the suspected AI-generated content was overly "sugarcoated."
• Set 4: The majority correctly identified the AI-generated content. However, opinions varied on
the quality of each text. Some participants found the AI text more fascinating and creative, while
others felt it was overly simplified for its target audience. The use of uncommon emojis in the AI
text was noted, with participants feeling that such a choice was unlikely to be made by a human.
On the other hand, some participants described the human-written text as plain and dull, leading
them to believe it was AI-generated.

115
• Set 5: None of the participants correctly identified the AI-generated content. The human-authored
text's formal tone and structured approach misled them. The group felt that the first text's attempt
to sound overly professional, using complex jargon, made it seem distant and unreliable, leading
them to mistakenly attribute it to AI.

The participants identified several attributes as signs of A.I.-generated copy, including language
patterns and a perceived lack of depth and understanding, particularly for complex topics. They also
reported the absence of personalization and emotional resonance in A.I.-generated content. On the
other hand, human-generated copy was associated with more natural language, personalized
elements, and the ability to evoke emotional responses. The discussion acknowledged the potential
evolution of AI's capabilities but emphasized the enduring value of human creativity and emotional
insight.

After sharing the results, the discussion culminated in a significant shift in the participants'
perspectives on AI and human-generated copy. The participants began with the idea that AI would
produce overly complex and lengthy language and that human copy would be more engaging and
accessible. However, after participating in the experiment, the participants recognized that AI could
produce high-quality content, often challenging their initial assumptions. The realizations sparked a
shift in their perception of AI's capabilities yet underscored the vital role of human input in guiding
AI's performance.The participants recognized the importance of human elements like hashtags and
emojis and the intentional structure of content for reader ease. Despite AI's impressive capabilities,
the group did not express a heightened fear of AI takeover. They viewed AI as a powerful tool
aiding the copywriting process but needed to see it as a complete substitute for human copywriters.

116
7.5 Experiment’ Findings
The digital experiment attracted 514 respondents, segmented as follows:
• Age: 66.5% born between 1997-2010; 33.5% between 1981-1996.
• Gender: 69.6% females, 27.2% males, and 3.1% non-binary.
• Education: Predominantly high school diploma holders (43.3%), followed by bachelor's
(25%), Master's (14.3%), doctorate (5.5%), writing academy (2.9%), and others (5%).
• Occupation: 'Student' led the list (307 respondents), followed by 'Researcher' (46) and
'Content Creator' (27).

The initial segment gauged participants' social media engagement. A staggering 99% were active on
Instagram, 74.9% on Facebook, 66.1% on TikTok, 47.4% on Pinterest, 26.3% on Twitter, 24.4% on
LinkedIn and varying engagement on other platforms. A generational divide was evident, with older
participants leaning towards Facebook and Twitter and younger ones towards TikTok and Snapchat.
However, platforms like Instagram and Pinterest transcended generational boundaries. Notably,
85.8% of respondents followed brands on social media. Of the participants who indicated following
brands on social media, 12.7% reported encountering brand content at a frequency level of 1 (the
least frequent). This was followed by 25.8% at a frequency level of 2, 34.8% at level 3, 21.7% at
level 4, and 5% at level 5 (the most frequent). The following sector aimed to look into how people
interacted with A.I. Most participants had a level 3 out of 5 acquaintance with artificial intelligence.
Their confidence in their ability to recognize artificial intelligence-generated content also had a
level 3 out of 5.

The core experiment presented five couples of human-crafted and AI-generated text created by
ChatGPT with the GPT-3.5 language model. Rigorous quality assurance ensured content parity. The
results revealed varied abilities among participants to trace content origin, with significant
deviations in some text sets. In these graphics, the results of the experiment are illustrated:

117
 Figure 15: Correct Answers to Set 1

50,39% 49,61%

Correct answer Wrong answer

Figure 16: Correct Answers to Set 2

37,89%

62,11%

Correct answer Wrong answer

118
 Figure 17: Correct Answers to Set 3

40,04%

59,96%

Correct answer Wrong answer

Figure 18: Correct Answers to Set 4

38,67%

61,33%

Correct answer Wrong answer

119
 Figure 19: Correct Answers to Set 5

30,47%

69,53%

Correct Answer Wrong answer

7.6 Analysis of Correct Answers Based on Educational Degree:

Utilizing SPSS for in-depth data analysis, we delved into the correlation between the educational
backgrounds of participants and their proficiency in identifying the origin of content. The
comprehensive analysis encompassed the following educational qualifications:

• Bachelor's Degree Holders:

◦ Total Participants: 127
◦ Total Correct Answers: 660
◦ Average Correct Answers per Participant: Approximately 5.20
• Doctorate/Ph.D. Holders:
◦ Total Participants: 28
◦ Total Correct Answers: 136
◦ Average Correct Answers per Participant: Approximately 4.86
• High School Diploma or Equivalent:
◦ Total Participants: 240

120
 ◦ Total Correct Answers: 1226
◦ Average Correct Answers per Participant: Approximately 5.11
• Master's Degree Holders:
◦ Total Participants: 72
◦ Total Correct Answers: 408
◦ Average Correct Answers per Participant: Approximately 5.67
• Post-graduate Degree Holders:
◦ Total Participants: 1
◦ Total Correct Answers: 4
◦ Average Correct Answers per Participant: 4.00
• Secondary School or Equivalent:
◦ Total Participants: 30
◦ Total Correct Answers: 146
◦ Average Correct Answers per Participant: Approximately 4.87
• Writing Academy Graduates:
◦ Total Participants: 15
◦ Total Correct Answers: 120
◦ Average Correct Answers per Participant: 8.00

7.8 Analysis of Correct Answers Based on Generation

This section delves into the correlation between the age of the participants and their ability to
identify the origin of content. As explicated in the following graphic, there is not a critical
difference between the two generation analysed:
Millenial participants were 171 and had a total of 934 correct answers, resulting in an average of
approximately 5.4 correct answers per individual.
Generation Z participants were 342 and had a total of 1766 correct answers, resulting in an average
of approximately 5.1 correct answers per individual.

121
 Figure 20: Correct Answers x Generations

49%
51%

Millenials (1981 -1996)

Generation Z (1997 - 2010)

7.9 Analysis of Correct Answers Based on Brand Content Frequency

This section delves into the correlation between the frequency with which participants encounter
brand content on social media and their ability to identify the origin of content:
• Frequency 1: Among the 58 participants in this category, 322 correct answers were given,
resulting in an average of approximately 5.55 correct answers per individual.
• Frequency 2: For this frequency, 117 participants provided 564 correct answers, averaging
approximately 4.82 correct answers per participant.
• Frequency 3: With 158 participants, there were 820 correct answers, leading to an average
of approximately 5.19 correct answers per individual.
• Frequency 4: From the 99 participants in this category, 564 correct answers were recorded,
translating to an average of approximately 5.70 correct answers per participant.
• Frequency 5: This group, consisting of 23 participants, registered 134 correct answers, with
an average of approximately 5.83 correct answers per individual.

122
 Figure 21: Correct Answers per Frequency

6
Average correct answer per person

5,697 5,826
5,552
5,19
4,5 4,821

1,5

0
1 2 3 4 5
How frequent the participant encounter written contents on social media

7.10 Analysis of Correct Answers Based on Familiarity with A.I.

This segment focuses on the potential correlation between participants' self-reported familiarity
with A.I. and their proficiency in distinguishing content origins:
• Familiarity Level 1: Among the 49 participants, 246 correct answers were recorded,
averaging approximately 5.02 correct answers per individual.
• Familiarity Level 2: This level had 130 participants who collectively provided 658 correct
answers, resulting in an average of approximately 5.06 correct answers per participant.
• Familiarity Level 3: With 177 participants, there were 850 correct answers, translating to
an average of approximately 4.80 correct answers per individual.
• Familiarity Level 4: From the 117 participants in this category, 712 correct answers were
given, leading to an average of approximately 6.09 correct answers per participant.
• Familiarity Level 5 (Most Familiar): This group consisted of 40 participants who
registered 234 correct answers, with an average of approximately 5.85 correct answers per
individual.

123
 Figure 22: Correct Answers per Familiarity

7
Average correct answer per person

6,085 5,85
5,25
5,02 5,062 4,802
3,5

1,75

0
1 2 3 4 5
Familiarity with A.I

124
 7.11 Analysis of Correct Answers Based on Confidence in Recognizing A.I. Content
This section provides a comprehensive breakdown based on participants' self-assessed confidence
levels in recognizing AI-generated content:
• Confidence Level 1: Out of 114 participants, there were 598 correct answers, resulting in an
average of approximately 5.25 correct answers per individual.
• Confidence Level 2: This level had 137 participants who collectively provided 684 correct
answers, averaging approximately 4.99 correct answers per participant.
• Confidence Level 3: With 148 participants in this category, 782 correct answers were given,
translating to an average of approximately 5.28 correct answers per individual.
• Confidence Level 4: Among the 97 participants, 550 correct answers were recorded, leading
to an average of approximately 5.67 correct answers per participant.
• Confidence Level 5: This group, consisting of 17 participants, registered 86 correct
answers, with an average of approximately 5.06 correct answers per individual.

Figure 23: Correct Answers per Confidence

6
Average correct answer per person

5,7
5,67

5,4

5,246 5,284
5,1
5,059
4,993
4,8
1 2 3 4 5
Confidence in recognising A.I. written content

125
8. Discussion:

8.1 Discussion of the Results of the Focus Groups:

8.2 General observations
Examining both focus group transcripts, shared observations and distinct perspectives become
evident. These insights form the basis of the present analysis, highlighting key patterns,
commonalities, and areas of divergence among the participants' views.
Shared Observations
1. Perceptions of A.I. Competence: All participants' initial perceptions tended to
underestimate A.I.'s capabilities in generating high-quality content. At the beginning of the
discussion, 10 participants out of 12 were optimistic about identifying AI-generated content
quickly. Participants were initially inclined to associate AI-generated text with overly formal
language and stringent grammatical precision. However, as the experiment unfolded and the
participants were exposed to the AI-generated content, a palpable shift in their perceptions
was observed. The revelation phase was particularly enlightening, as participants discovered
that specific texts they were confident were human-written were, in fact, products of A.I.
This realization challenged and reshaped their initial assumptions, leading to a more
nuanced understanding of A.I.'s capabilities in copywriting. In some cases, the surprise and
disbelief at the results underscored the participants' initial underestimation of A.I.'s
capabilities. They began to recognize that A.I. could produce content that matched and
sometimes even surpassed the quality of human composition.
2. Distinctly Human Characteristics: Both focus groups were inclined to identify what they
believed were uniquely 'human' traits within the texts. Elements such as a well-thought-out
structure aimed at enhancing readability, wordplay, and employing more intense punctuation
were frequently cited. These characteristics were almost unanimously considered indicators
of a 'human touch.' The discussions also touched upon the structure of the content. For
instance, dense blocks of text without clear paragraph breaks were sometimes seen as
indicative of machine generation, as humans tend to structure their writing to enhance
readability. The discussions also ventured into the realm of emojis and hashtags. Participants
from both groups felt that how emojis and hashtags were used could offer clues about the

126
 text's origin. For instance, using uncommon or excessive emojis was often attributed to A.I.
Moreover, the capacity to convey genuine sentiment and depth in writing was deemed a
distinctly human trait. Participants felt that while A.I. could mimic human writing styles, it
could not replicate the genuine emotion and depth humans naturally infuse into their writing.
Another sign often associated with A.I. was the text to "sugarcoated".
3. Complexity of Language and Readability: A recurring theme in both focus groups was the
discussion around the complexity of language and its readability. Participants often
associated mechanical precision and strict adherence to grammatical norms with A.I.-
generated content. In contrast, the presence of linguistic quirks, catchy calls to action,
wordplays, and even occasional errors were seen as hallmarks of human writing. Some
participants pointed out that while A.I. might construct grammatically perfect sentences, it
might need help replicating the natural flow and rhythm of human writing.
4. Dependence of A.I. on Human Input: Despite acknowledging A.I.'s unexpected skills, the
consensus from both groups was firm in asserting that the efficacy of A.I. is as good as the
quality and range of data fed to it. Overall, just three participants needed to be made aware
of how a language model operates and that it has been trained to rely on a limited number of
data. On the other hand, just one participant was informed enough to understand the
differences between various language models and their capacities and hypothesize how they
could evolve in the future. It was acknowledged that while A.I. offers significant
advantages, it can only function at its optimal capacity with human oversight, empathy, and
creativity.

127
8.2 Contrasting Perspectives
The differences in interpretation can be attributed to each participant's unique set of experiences,
varying levels of familiarity with AI, and their perspectives on the role of technology within society.
Participants who are more receptive to technological advancements may be due to positive
experiences or inherent beliefs, while others may exercise caution when considering such
developments. These elements shape individual perceptions and interpretations, resulting in the
noted differences in opinions.The analysis of both focus group discussions illuminates a shared
understanding of AI's abilities and potential within copywriting. It simultaneously reveals a
spectrum of opinions regarding the implications of these capabilities. This indicates that while some
aspects of AI in copywriting have reached a consensus, the broader conversation is far from
definitive and continues to represent many perspectives.

8.3 Differences between Generation Z and Millennials

Throughout the focus groups, specific trends and differences were apparent in the responses and
attitudes of Gen Z and Millennial participants. Both generations showed a significant familiarity
with technology, particularly artificial intelligence and language models, with a relatively equal
number of individuals using AI tools across both generations.

However, a clear distinction emerged regarding social media usage and brand engagement. Gen Z
participants, often hailed as digital natives, consistently demonstrated high engagement with brands
on social media. They showed a thoughtful understanding of brands' marketing strategies and the
tone of voice used in their communications, even participants not working in marketing-related
fields. They also showed high awareness and interaction with social media platforms for brand
communication. In contrast, Millennial participants tended towards a more personal use of social
media. While they use these platforms frequently, their engagement was more directed towards
keeping up with friends rather than following brands.

Attitudes towards AI integration in creative works also indicated a generational divergence. Gen Z
participants, possibly shaped by their early exposure to rapidly advancing technologies, exhibited
high optimism and openness towards integrating AI in creative works. They believed AI could be

128
used beneficially in creative fields without threatening human roles in these sectors. On the other
hand, while Millennials' views on this subject were not explicitly laid out in the discussion, they
might exhibit a more cautious or ambivalent attitude given their delayed exposure to advanced
technology.

Lastly, examining textual elements, such as emojis, hashtags, sentence structure, and punctuation,
reflected a distinct generational divide. Gen Z participants frequently focused on modern, social-
media-oriented features like hashtags and emojis, viewing them as critical textual cues to discern
the origin of a text. In contrast, Millennials focused more on traditional aspects of writing, such as
the complexity of sentence structure and punctuation. This difference underscores a generational
shift in the understanding and interpretation of text, with Gen Z prioritizing the newer elements
indicative of the digital age and Millennials emphasizing more conventional, established elements
of writing.

8.4 Overall Analysis of the Focus Groups

In exploring A.I.'s transformative potential within copywriting, the focus groups provided
invaluable insights into human perceptions of A.I.-generated content. The main goal of the focus
groups was to retrieve qualitative data that would explain the prejudice, ideas, and feelings that
social media users encounter when related to A.I.

The participants, spanning diverse backgrounds and age groups, initially underestimated A.I.'s
prowess in producing high-quality content. However, a tangible shift in perceptions emerged as the
discussion progressed, challenging their initial assumptions. Both groups identified uniquely
'human' traits in texts, such as genuine sentiment, depth, and linguistic quirks. They also highlighted
the importance of human oversight in guiding A.I.'s performance. Notably, generational differences
between Gen Z and Millennials surfaced, particularly in their engagement with brands on social
media and their interpretations of textual elements. While both groups recognized A.I.'s growing
capabilities, they unanimously emphasized the irreplaceable value of human creativity and
emotional depth in copywriting.

129
This study underscores the importance of understanding potential biases and perceptions people
hold about AI-generated content, as these perceptions shape the evolving discourse on A.I.'s role in
creative fields. The results of the focus group highlight how, even if not always able to correct and
identify A.I. generated text, all the participants expressed strong bias regarding the use of A.I., and
whenever the text felt too "fake" or "robotic" they showed a feeling of dislike. These results
emphasized that, at the actual stage, LLMs can not replace the work of a human copywriter, as all
the participants looked for "the human touch" in the texts.

8.5 Discussion of the Results of the Experiment:

8.6 Analysis of Correct Answers in Relation to Educational Highest Completed Level
In the complicated task of distinguishing AI-generated content from human-crafted content, the
influence of one's educational background emerges as a defining factor. Our comprehensive data
analysis reveals compelling data: the performance of Writing Academy Graduates is particularly
illuminating, with an average of 8 correct answers per participant; this group's proficiency in
identifying content origins eclipses that of all other educational categories. This is not a mere
statistical anomaly but proof of the depth of understanding and expertise that specialized writing
training confers.
Contrast this with the performance of participants from other educational backgrounds. Whether
holding a Bachelor's, Master's or even a Doctorate, the differences in their ability to discern content
origins were marginal. Despite their varied academic levels, these groups hovered around similar
averages, ranging from a medium of 4.86 to 5.67 correct answers per person. The data suggests that
while general academic achievements offer a breadth of knowledge, they must equip individuals
with the nuanced language understanding that specialized writing education does.
This observation underscores an insight: An academic journey centered on writing, creative or
copywriting, cultivates a unique linguistic sensitivity. Such training enables individuals to perceive
the subtle intricacies and characteristics distinguishing human touch from AI-generated content.
While A.I. continues to make strides in content generation, an innate quality to human writing
resonates distinctly with those trained to recognize it.
In conclusion, while all educational backgrounds bring their strengths to the table, the specialized
focus on writing sets individuals apart in discerning A.I. from human-crafted content.

130
8.7 Analysis of Correct Answers in Relation to Generation
The correlation between participants' age and ability to discern the origin of content is intriguing.
As illustrated in the data, there is a subtle difference in the average correct answers between
Millennial and Generation Z participants. Specifically, Millennials averaged approximately 5.4
correct answers, while Generation Z averaged around 5.1 correct answers.
Gen Z, often termed "digital natives," are individuals inherently familiar with the digital realm,
encompassing computers, the internet, phones, and video games. This generation has grown up with
smartphones and is adept at using multiple devices. Their approach to technology is intuitive rather
than instructive; they prefer hands-on exploration over reading manuals. This contrasts with
previous generations who relied on encyclopedias for information. Interestingly, Millennials and
Gen Z spend a similar amount of time online, averaging over 7 hours daily. However, Gen Z's
attachment to technology, especially smartphones, promotes a unique form of omnipresent social
engagement (Serbanescu, 2022).
The rapid evolution of technology and its integration into daily life has led to speculations about its
impact on the human brain. Some argue that the constant exposure to digital platforms might
reshape our neural pathways, suggesting a potential "epigenetic" modification (Negroponte, 1995 as
quoted by Serbanescu, 2022). While this might sound far-fetched, the idea that our brains adapt to
our experiences is not new. The continuous interaction with digital platforms might enhance
specific cognitive skills in Gen Z, such as multitasking, nonlinear thinking, and the ability to
connect disparate pieces of information (Serbanescu, 2022).
Considering the distinct social media usage patterns between Millennials and Gen Z, which are also
evident in the social media indicated by each generation in this research, it is intriguing that their
ability to discern AI-generated content is relatively similar. This observation underscores the
complexity of factors influencing discernment. It suggests that mere exposure to technology or
specific platforms is not the sole determinant of one's ability to identify AI-generated content.

8.8 Analysis of Correct Answers in Relation to Brand Content Exposure

The data suggests that the frequency of brand content exposure on social media does not directly
correlate with the ability to discern AI-generated content. Increased exposure does not necessarily

131
enhance discernment abilities. This indicates that the quality of engagement, rather than the volume
of content encountered, might be more influential in shaping discernment skills.

8.9 Analysis of Correct Answers in Relation to A.I. Familiarity

Data patterns reveal that more profound A.I. familiarity sometimes equates to heightened
discernment. Participants who expressed a familiarity between levels 1 and 3 had almost the same
results, but those with results 4 and 5 had a slightly higher number of correct replies. Given the
evolving nature of AI-generated content, these results highlight that perceived familiarity is only
sometimes equivalent to reality and that continuous adaptation and recalibration of understanding
are essential.

8.10 Analysis of Correct Answers in Relation to Confidence and Discernment

While one might expect participants with the highest confidence to excel in discernment, the results
challenge this assumption. Those expressing maximum confidence did not significantly outperform
their less confident counterparts. This disparity suggests that factors beyond A.I. understanding,
such as ego or self-perception, might influence expressed confidence. The data emphasizes the need
to differentiate between genuine expertise and perceived self-efficacy in discerning A.I.-generated
content.

8.11 Overall Results of the Experiment

The findings of this experiment present an intriguing insight. Aside from the notable exception of
students or alums of writing academies, there appear to be no definitive variables that significantly
influence an individual's capacity to recognize AI-generated content. While various factors such as
educational background, generational differences, brand content exposure, A.I. familiarity, and
confidence levels were meticulously analyzed, none exhibited the pronounced impact that
specialized writing education did. This underscores the unique skill a focused writing curriculum
imparts in discerning the nuances between human and AI-crafted content. In a landscape where
A.I.'s capabilities are ever-evolving, the results emphasize the importance of specialized training
while highlighting the multifaceted nature of human discernment that remains largely unaffected by
other variables.

132
In conclusion, this experiment confirmed that the capacity of recognize A.I. is completely
subjective, and it's innate in subjects with a particular attention to writing, but can also be trained
with an adequate academic journey.

8.12 Triangulation of Results

Triangulation, a method used to validate research findings by cross-verifying them with two or
more methods, is a crucial aspect of our study (Heale & Forbes, 2013; O'Cathain et al., 2010;
Tashakkori & Creswell, 2007). By combining the quantitative results from the experiment with the
qualitative insights from the focus groups, we aimed to provide a comprehensive understanding of
participants' abilities to discern AI-generated content. This approach not only strengthened the
validity of our findings but also shed light on the complexities and nuances that a single method
might overlook (Heale & Forbes, 2013; O'Cathain et al., 2010; Tashakkori & Creswell, 2007).

8.13 Comparative Analysis of Experimental Results and Focus Group Insights

The combined data from the experiment and focus group discussions offer a complete
understanding of participants' abilities and motivations to discern AI-generated content. While the
experimental results provided a quantitative overview, the focus groups added depth by uncovering
motivations, biases, and thought processes.

The experimental results and the insights from the focus groups converged on a particular
observation: participants consistently underestimated the capabilities of A.I. before participating in
the experiment. During the beginning of the focus group discussions, a significant majority
expressed confidence in their capacity to discern A.I.-generated content from human-generated
content. This sentiment was further echoed in the survey results, where a notable portion of
participants rated their confidence level as a three on a 1 to 5 scale. However, a strong contrast
emerged with their performance, revealing a gap between their perceived abilities and actual
discernment skills. The post-discussion segments of the focus groups were particularly
enlightening. They illuminated the depth and breadth of misconceptions many held about AI-
generated content. Participants seemed to have a stereotypical picture of A.I.-generated texts,
expecting A.I. outputs to be overtly robotic, lacking nuance, or following a predictable pattern.

133
Nevertheless, the reality of A.I.-generated content is far more complex. It reflects not just
algorithms and computations but also the human touch, how the A.I. is trained, the prompts given to
the program, and the desired outcome the human operators aim for. One participant from the first
focus group captured this evolving realization succinctly, stating, "I used to think that robots would
take over every career, in a very sci-fy vision of the future, but now I am much more scared of
people that can use A.I., they will take over all our jobs so easily." This profound shift in
perspective underscores the duality of A.I.: It is not just a tool but an extension of human intent.
Another participant further emphasized this transformation in perception, moving from viewing A.I.
as a looming threat to recognizing its potential as a formidable tool. When harnessed correctly, A.I.
can replicate human tasks and amplify, refine, and elevate them, offering possibilities previously
deemed unattainable.

Combined with the results from both methodologies, they provide a comprehensive understanding
of generational differences in discernment abilities. Notably, there is a subtle distinction in how
Millennials and Generation Z participants interact with and perceive digital content. Gen Z, often
called "digital natives," have grown immersed in the digital environment. Their engagement with
technology is instinctual, leaning towards hands-on exploration rather than relying on structured
instructions. This contrasts with Millennials, who experienced a transition from traditional
information sources, like encyclopedias, to the vast digital resources available today (Serbanescu,
2022). While foundational differences exist in how each generation was introduced to the digital
world, both spend a similar amount of time online. Both generations also demonstrate a deep
understanding of social media communication. However, the methodologies highlighted a
generational divergence in platform preference. However, even with these variations in digital
interaction and platform choices, the ability of both generations to identify AI-generated content is
remarkably consistent. This highlights that mere technological exposure is not the sole determinant
of discernment abilities. The focus group discussions further enriched this narrative. Both
generations exhibited significant familiarity with A.I. and language models. However, clear
distinctions emerged in their engagement patterns. Gen Z, with their inherent digital fluency,
displayed a keen understanding of brand communications on social media. They also showcased a
heightened awareness of modern textual elements like hashtags and emojis, viewing them as critical

134
cues when recognizing human written content. On the other hand, millennials leaned towards
personal connections on social media and emphasized traditional textual elements like sentence
structures and punctuation. This difference underscores a generational shift in textual interpretation.

Another point of convergence between the focus groups and survey results was the shared optimism
and apprehension regarding A.I.'s role in creative endeavors. Gen Z's optimism about A.I.'s potential
in creative fields was palpable, viewing it as a complementary tool rather than a threat. In contrast,
due to their phased introduction to advanced technology, Millennials might harbor more cautious or
ambivalent views. While both generations are deeply embedded in the digital age, their interactions,
interpretations, and attitudes towards technology and A.I. showcase intriguing differences, painting
a multifaceted picture of the evolving human-tech relationship.

The combination of the two methods, one quantitative and one qualitative, underscores the
importance of the motivations and the ideas behind participants' discernment. Additionally, the
focus groups revealed participants' grasp of A.I.'s reliance on human input and their ability to
recognize human-like characteristics in content. Such insights were less pronounced in the
experimental data, underscoring the value of qualitative research in understanding human
cognition's complexities.

While both methodologies were complementary, they also had points of divergence. For instance,
the quantitative data emphasized the heightened discernment abilities of those with specialized
writing training. However, the focus groups lacked participants with an academic emphasis on
writing, which might have offered more profound insights into their unique perspectives. However,
irrespective of their educational background, the focus groups consistently revealed an overarching
underestimation of A.I.'s capabilities. This indicates that while specialized training provides
discernment tools, there is a broader challenge in comprehending A.I.'s evolving capabilities.

135
9. Conclusion
In the rapidly evolving digital landscape, integrating technology into different sectors has become
essential; everything is changing as fast as technological progress, from the legal field to the
marketing and advertising world. Artificial intelligence is just the last of a long series of inventions
that led to a revolution in every aspect of life, private and professional.Writing proposes a particular
challenge among all these sectors due to its intrinsic connection with human communication and
expression. This research embarked on an exhaustive exploration of the prospective influence of
A.I. on the sphere of copywriting, aiming to provide a holistic understanding of its transformative
capabilities.

The primary objective of this study was to delve into the intricate relationship between A.I. and
copywriting. With the advent of sophisticated A.I. language models, there is an inevitable
transformation in the written content production field. However, it is still complex to understand the
volume of these technologies' impact in the future. This research's driving question was to unravel
the multifaceted consequences of A.I.'s integration into copywriting. The secondary investigative
question provided a pragmatic dimension to the research. By examining individuals' ability to
differentiate between human-crafted and A.I.-generated content, this research illustrated the
perception of A.I. and A.I.-generated content and the challenges associated with trust, authenticity,
and the broader ethical considerations in A.I.'s role within copywriting. Understanding the pre-
concepts that social media users have on A.I.-generated content and which patterns they are looking
for to identify a human-created text is fundamental to understanding if the destiny of A.I. language
model is to take over every job related to writing or if it is going to become a powerful tool for
human creators.

The study's primary focus on millennials and Generation Z was strategic. These digital natives, with
their intrinsic connection to online platforms, provided invaluable insights into the evolving
landscape of copywriting (Serbanescu, 2022). Their digital fluency, combined with their unique
perspectives on technology, offered a fresh lens through which the implications of AI-powered
copywriting could be viewed. The research's sampling structure was meticulously designed to

136
capture various viewpoints. Ensuring representation from diverse demographics enriched the study
with multifaceted experiences, providing a more comprehensive understanding of the topic.

Triangulation, a technique that cross-checks research findings using multiple methods, was a key
component of our study. The research merged quantitative data from experiments to reach a more
significant sample, with qualitative feedback from focus groups to gain a deeper insight into how
participants discern AI-generated content, their ideas, and prejudice on A.I.. This dual-method
approach validated our results and highlighted intricate details that might have been missed using
just one method.

Participants' consistent underestimation of A.I.'s capabilities was a salient observation across both
methodologies. Despite entering the study confidently in discerning AI-generated content, the
outcomes highlighted a discernible gap between participants' perceived competencies and their
actual discernment abilities.

One of the most striking research revelations found in the experiment's results was the exceptional
ability of individuals with specialized writing training to recognize A.I. Their profound academic
grounding in writing, both in copywriting and creative domains, significantly influenced their
ability to recognize A.I.-generated content.

In contrast, other variables like age, A.I. familiarity, and brand content exposure frequency on social
media had a surprisingly minimal impact on discernment abilities. Despite their distinct digital
habits, the subtle differences in discernment between Millennials and Generation Z further
emphasized the general difficulty of any social media user distinguishing A.I.-generated content.

The focus groups emphasized the stereotypes associated with A.I.-generated content: a high tone of
voice, particularly in social media content, "sugarcoated" messages, repetition, and an unusual
selection of emojis. While the experiment did not delve into the reasons behind each participant's
response, the outcomes closely mirrored those of the focus groups. Thus, it is plausible to
hypothesize that these stereotypes were also prevalent in the minds of the experiment participants.

137
These results highlight how, unless the subject does not have a specific writing academic journey,
all the social media users have a different understanding and perception of A.I. content. Therefore,
both methodology's results suggest that if A.I. has the potential to become an incredible tool for
copywriters, it is improbable that it will completely replace their jobs, as consumers may be able to
recognize the difference and search for a human touch in the content they came across.

9.1 Implications of the Research

This research looked at the complicated relationship between A.I. and copywriting, finding that
while A.I. language models are transforming content production, the human touch remains
indispensable, as social media users associate A.I. with particular traits of written content. This is an
essential discovery for everyone who works in copywriting, but much more so for anyone who
works with social media. The research results demonstrate how human creativity, thinking, and
direction are irreplaceable. This shows that while A.I. can effectively supplement copywriters' work,
more training and implementation is required to substitute human copywriters.

9.2 Limitations
Like any research, this study has its limitations. A limitation of the focus group was the absence of
participants from a creative writing academy, which could have enriched the insights and further
validated the experimental results due to their specialized training in writing. While practical, the
study's reliance on convenience sampling might not have delivered a fully representative sample,
introducing potential biases and limiting the generalizability of the findings. Conducting the
research in a controlled environment might not replicate the nuances of real-world conditions,
suggesting that participants' responses could differ in more natural settings. Time constraints
inherent to the study design limited the depth of exploration into specific themes or the breadth of
participants' perspectives. The survey's structured nature restricted the depth of understanding
regarding the reasons behind participants' responses, and a more open-ended approach might have
offered richer insights. Cultural considerations were also overlooked; different cultures may have
varied perceptions and approaches to A.I., which could influence discernment abilities and attitudes.
Lastly, by focusing primarily on Millennials and Generation Z, the study inadvertently excluded the

138
perspectives of older generations, potentially missing out on their unique insights and experiences
with A.I.

9.3 Suggestions for Future Research

Concerning A.I., the domain of copywriting remains relatively uncharted in academic research.
While many studies focus on A.I.'s broader implications on content creation, the nuanced intricacies
specific to copywriting should be noticed. Furthermore, a significant lacuna persists in
understanding public perception of AI-generated copy. Many studies might touch upon the
capabilities of A.I. in content creation. However, only some, if any, delve deep into discerning how
audiences perceive and react to AI-crafted content versus human-generated copy. This absence of
focused research not only underscores the novelty of our study but also emphasizes the pressing
need for more granular investigations in this domain. This negligence is particularly notable, given
the transformative potential of A.I. in reshaping the copywriting landscape. This research offers a
first overview of social media users' perception of A.I.-generated content and, therefore, the volume
of LLMs' impact on the copywriting field. With its unique focus, this research underscores the
pressing need for more targeted investigations in this domain. The absence of previous research and
experiments about the ability of social media users to recognize A.I. accentuates the originality and
importance of this study, emphasizing the need for future research endeavors to delve deeper into
this niche area.

In conclusion, this research is the first step into fully understanding A.I.'s impact on copywriting.
However, as A.I. technologies continue to evolve, considering the newest technologies, it is
essential to keep exploring the topic. The results of the research responded to the initial research
question, emphasizing the irreplaceable value of the human touch in content creation and
highlighting that social media users do not have the ability to recognize A.I. consistently; however,
they can recognize patterns utilized by LLM for creating written content. This proves the
potentiality of A.I. but also highlights how it is not enough to substitute a human copywriter at the
current stage of development.

139
10. Bibliography

• Aaker, D. (2014). Aaker on branding: 20 principles that drive success. Morgan James Publishing.
• Aaker, D. A. (1997). Dimensions of brand personality. Journal of Marketing Research, 34(3),
347-356.
• Aaker, J. L. (1999). The malleable self: The role of self-expression in persuasion. Journal of
marketing research, 36(1), 45-57.
• About. (n.d.). https://openai.com/about
• Akhter, R. (2018). Impact of copywriting in marketing communication. Journal of Advertising,
46(3), 387-398.
• Albrighton, T. (2013). The abc of copywriting. Norwich: ABC Business Communications Ltd.
• Alekseeva, L., Azar, J., Gine, M., Samila, S., & Taska, B. (2020). The demand for AI skills in the
labor market. Labour economics, 71, 102002
• Anderson, N., Belavý, D. L., Perle, S. M., Hendricks, S., Hespanhol, L., Verhagen, E., & Memon,
A. R. (2023). AI did not write this manuscript, or did it? Can we trick the AI text detector into
generated texts? The potential future of ChatGPT and AI in Sports & Exercise Medicine
manuscript generation. BMJ Open Sport and Exercise Medicine, 9(1), e001568. https://doi.org/
10.1136/bmjsem-2023-001568
• Applegate, E. (2015). Strategic Copywriting: How to Create Effective Advertising. Rowman &
Littlefield.
• Arsenijevic, U., & Jovic, M. (2019). Artificial Intelligence Marketing: Chatbots. 2019
International Conference on Artificial Intelligence: Applications and Innovations (IC-AIAI).
https://doi.org/10.1109/ic-aiai48757.2019.00010
• Azoulay, A., & Kapferer, J. N. (2003). Do brand personality scales really measure brand
personality?. Journal of brand management, 11, 143-155.
• Ballardini, R. M., He, K., & Roos, T. (2019). AI-generated content: authorship and inventorship in
the age of artificial intelligence. In Edward Elgar Publishing eBooks. https://doi.org/
10.4337/9781788119900.00015
• Baltes, L. P. (2017). Marketing technology (Martech)–the most important dimension of online
marketing. Bulletin of the Transilvania University of Brasov. Series V: Economic Sciences, 43-48.

140
• Beard, F. (2017). The Ancient History of Advertising: Insights and Implications for Practitioners.
Journal of Advertising Research, 57(3), 239–244. https://doi.org/10.2501/jar-2017-033
• Bechter, C., Farinelli, G., Daniel, R. D., & Frey, M. (2016). Advertising between archetype and
brand personality. Administrative Sciences, 6(2), 5.
• Berger, J., & Milkman, K. L. (2012). What makes online content viral? Journal of Marketing
Research, 49(2), 192-205.
• Bettig, R. V. (2018). Copyrighting culture: The political economy of intellectual property.
Routledge.
• Blynova, N., & Kyrylova, O. (2018). Copywriting as an activity of online content creation.
Communications and Communicative Technologies.
• Boerman, S. C., Kruikemeier, S., & Zuiderveen Borgesius, F. J. (2017). Online behavioral
advertising: A literature review and research agenda. Journal of Advertising, 46(3), 363-376.
• Bond, C., Ferraro, C., Luxton, S., & Sands, S. (2010, November). Social media advertising: An
investigation of consumer perceptions, attitudes, and preferences for engagement. In Proceedings
of the Australian and New Zealand Marketing Academy (ANZMAC) Conference (pp. 1-7).
• Brauner, A. S., & Gymnasium, M. G. (2006) Runaround by Isaac Asimov and the significance of
the Three Laws of Robotics in today!s world.
• Brown, H., Lee, K., Mireshghallah, F., Shokri, R., & Tramèr, F. (2022). What Does it Mean for a
Language Model to Preserve Privacy? 2022 ACM Conference on Fairness, Accountability, and
Transparency. https://doi.org/10.1145/3531146.3534642
• Buchanan, B. G. (2005, December 15). A (Very) Brief History of Artificial Intelligence | AI
Magazine. https://ojs.aaai.org//index.php/aimagazine/article/view/1848
• Calo, R. (2017). Artificial intelligence policy: a primer and roadmap. UCDL Rev.
• Campbell, C., Sands, S., Ferraro, C., Tsao, H.Y.J., & Mavrommatis, A. (2020). From data to
action: How marketers can leverage AI. Business Horizons, 63(2), 227-243.
• Cao, Y. (2023, March 7). A Comprehensive Survey of AI-Generated Content (AIGC): A History
of Generative AI from GAN to ChatGPT. arXiv.org. https://arxiv.org/abs/2303.04226
• Carrozzo, R. (2023). OPEN? AI: GPT-4 e la sfida etica dell!Intelligenza Artificiale. Il Sole 24
Ore. https://www.ilsole24ore.com/art/open-ai-gpt-4-e-sfida-etica-intelligenza-artificiale-
AE1Y5SID?refresh_ce

141
• Castellanos-Gomez, A. (2023). Good Practices for Scientific Article Writing with ChatGPT and
Other Artificial Intelligence Language Models. Nanomanufacturing, 3(2), 135–138. https://
doi.org/10.3390/nanomanufacturing3020009
• Castelli, L., Vanzetto, K., Sherman, S. J., & Arcuri, L. (2001). The explicit and implicit perception
of in-group members who use stereotypes: Blatant rejection but subtle conformity. Journal of
Experimental Social Psychology, 37(5), 419–426.
Chen, F., Yue, X., Yang, X., & Ge, T. (2014). Study on classification of personality-based brand
archetype from the perspective of the internet. China Communications, 11(7), 153-160.
• Chintalapati, S., & Pandey, S. K. (2022). Artificial intelligence in marketing: A systematic
literature review. International Journal of Market Research, 64(1), 38–68. https://doi.org/
10.1177/14707853211018428
• Chiu, K. (2021, March 23). Detecting Hate Speech with GPT-3. arXiv.org. https://arxiv.org/abs/
2103.12407
• Cialdini, R. B., (2007). Influence: The psychology of persuasion (Vol. 55, p. 339). New York:
Collins.
• Cizmeci, F. (2015). The effect of digital marketing communication tools to create brand
awareness by housing companies. Megaron. https://doi.org/10.5505/megaron.2015.73745
• Clark, E., Ross, A. S., Tan, C., Ji, Y., & Smith, N. A. (2018, March). Creative writing with a
machine in the loop: Case studies on slogans and stories. In 23rd International Conference on
Intelligent User Interfaces (pp. 329-340).
• Clerwall, C. (2014). Enter the robot journalist: Users!"perceptions of automated content.
Journalism Practice, 8(5), 519–531.
• Corrales, M., Fenwick, M., & Haapio, H. (2019). Digital Technologies, Legal Design and the
Future of the Legal Profession. Legal Tech, Smart Contracts and Blockchain, 1–15. https://
doi.org/10.1007/978-981-13-6086-2_1
• Creswell, J. W. (2014). A concise introduction to mixed methods research. SAGE publications.
• Creswell, J. W., & Creswell, J. D. (2017). Research design: Qualitative, quantitative, and mixed
methods approaches. Sage publications.
• Creswell, J. W., Clark, V. L. P., Gutmann, M. L., & Hanson, W. E. (2003). ADVANCED
MIXED. Handbook of mixed methods in social & behavioral research, 209.

142
• Dale, R. (2021). GPT-3: What!s it good for? Natural Language Engineering, 27(1), 113–118.
https://doi.org/10.1017/s1351324920000601
• Dargham, M., Hachimi, H., & Boutalline, M. (2022). How AI is automating writing: The rise of
robot writers. https://doi.org/10.1109/icoa55659.2022.9934723
• Davenport, T. H., Guha, A., Grewal, D., & Bressgott, T. (2020). How artificial intelligence will
change the future of marketing. Journal of the Academy of Marketing Science, 48(1), 24–42.
https://doi.org/10.1007/s11747-019-00696-0
• Davis, J. L., Love, T. P., & Killen, G. (2018). Seriously funny: The political work of humor on
social media. New media & society, 20(10), 3898-3916.
• De Bruyn, A., Viswanathan, V., Beh, Y. S., Brock, J. K. U., & Von Wangenheim, F. (2020).
Artificial intelligence and marketing: Pitfalls and opportunities. Journal of Interactive Marketing,
51(1), 91-105.
• DeepL Nuova svolta nella qualità della traduzione basata sull’Intelligenza Artificiale. (2020,
February 6). https://www.deepl.com/it/blog/20200206
• Deng, J.; Lin, Y. The Benefits and Challenges of ChatGPT: An Overview. Front. Comput. Intell.
Syst. 2023, 2, 81–83.
• Dergaa, I., Chamari, K., Zmijewski, P., & Saad, H. B. (2023). From human writing to artificial
intelligence generated text: examining the prospects and potential threats of ChatGPT in academic
writing. Biology of Sport, 40(2), 615-622.
• Dobrev, D. (2012, October 3). A Definition of Artificial Intelligence. arXiv.org. https://arxiv.org/
abs/1210.1568
• Donahue, L. & Hajizadeh, F. (2019). Artificial Intelligence in cloud marketing. Artificial
Intelligence and machine learning for business for non-engineers, 77-88.
• Du Plessis, C. (2015). Brand storytelling: the case of Coca-Cola's journey corporate website.
• Du, S., & Xie, C. (2021). Paradoxes of artificial intelligence in consumer markets: Ethical
challenges and opportunities. Journal of Business Research, 129, 961-974.
• Duarte, A., & Neumaier, A. (2022). Chatvertising: how chatbots are shaping the future of
advertising. Comunicação Pública, 17(32).
• Elkins, K., & Chun, J. (2020). Can GPT-3 Pass a Writer!s Turing Test? Journal of Cultural
Analytics, 5(2). https://doi.org/10.22148/001c.17212

143
• Elliot, M. (2022). What is AI? In Architecture in the Age of Artificial Intelligence.
• Engelbart, D. C. (2023). Augmenting human intellect: A conceptual framework. In Augmented
Education in the Global Age (pp. 13-29). Routledge.
• Faber, M. A., & Mayer, J. D. (2009). Resonance to archetypes in media: There's some accounting
for taste. Journal of Research in Personality, 43(3), 307-322. https://doi.org/10.1016/
j.jrp.2008.11.003.
• Fan, A., Stock, P., Graham, B., Grave, E., Gribonval, R., Jegou, H., & Joulin, A. (2020). Training
with quantization noise for extreme model compression. arXiv preprint arXiv:2004.07320.
• Fennis, B. M., & Stroebe, W. (2010). The psychology of advertising. Psychology Press.
• Fetzer, J. H. (1990). What is Artificial Intelligence?. In Artificial Intelligence: Its Scope and
Limits
• Finzi, V. S. (2017). Storia della psicoanalisi. Autori, opere, teorie 1895–1990. Mondadori.
• Fitria, T. N. (2023, March). Artificial intelligence (AI) technology in OpenAI ChatGPT
application: A review of ChatGPT in writing English essay. In ELT Forum: Journal of English
Language Teaching (Vol. 12, No. 1, pp. 44-58).
• Floridi, L., & Chiriatti, M. (2020). GPT-3: Its Nature, Scope, Limits, and Consequences. Minds
and Machines, 30(4), 681–694. https://doi.org/10.1007/s11023-020-09548-1
• Frank, M. R., Autor, D. H., Bessen, J., Brynjolfsson, E., Cebrian, M., Deming, D., Feldman, M.
P., Groh, M., Lobo, J., Moro, E., Wang, D., Youn, H., & Rahwan, I. (2019). Toward understanding
the impact of artificial intelligence on labor. Proceedings of the National Academy of Sciences of
the United States of America, 116(14), 6531–6539. https://doi.org/10.1073/pnas.1900949116
• Ganassali, S., & Matysiewicz, J. (2021). Echoing the golden legends: Storytelling archetypes and
their impact on brand perceived value. Journal of Marketing Management, 37(5-6).
• Garnelo, M., & Shanahan, M. (2019). Reconciling deep learning with symbolic artificial
intelligence: representing objects and relations. Current Opinion in Behavioral Sciences.
• GPT-4. (n.d.). https://openai.com/research/gpt-4
• Graefe, A., Haim, M., Haarmann, B., & Brosius, H. B. (2018). Readers!"perception of computer-
generated news: Credibility, expertise, and readability. Journalism, 19(5), 595–610.
• Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: attitudes, self-esteem, and
stereotypes. Psychological Review, 102(1), 4–27.

144
• Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. (1998). Measuring individual differences in
implicit cognition: the implicit association test. Journal of Personality and Social Psychology,
74(6), 1464.
• Haenlein, M., & Kaplan, A. (2019). A Brief History of Artificial Intelligence: On the Past,
Present, and Future of Artificial Intelligence. California Management Review, 61(4), 5–14.
https://doi.org/10.1177/0008125619864925
• Haleem, A., Javaid, M., Qadri, M. A., Singh, R. P., & Suman, R. (2022). Artificial intelligence
(AI) applications for marketing: A literature-based study. International Journal of Intelligent
Networks, 3, 119–132. https://doi.org/10.1016/j.ijin.2022.08.005
• Hameed, M. A., Hassaballah, M., Hosney, M. E., & Alqahtani, A. (2022). An AI-Enabled Internet
of Things Based Autism Care System for Improving Cognitive Ability of Children with Autism
Spectrum Disorders. Computational Intelligence and Neuroscience, 2022, 1–12. https://doi.org/
10.1155/2022/2247675
• Hancock, J. T., Naaman, M., & Levy, K. (2020). AI-mediated communication: Definition,
research agenda, and ethical considerations. Journal of Computer-Mediated Communication,
25(1), 89-100.
• Handley, A. (2014). Everybody writes: your go-to guide to creating ridiculously good content.
John Wiley & Sons.
• Harper, G. (2010). On creative writing (Vol. 4). Multilingual Matters.
• Heale, R., & Forbes, D. (2013). Understanding triangulation in research. Evidence-Based
Nursing, 16(4), 98. https://doi.org/10.1136/eb-2013-101494
• Heaven, W. D. (2021, October 20). OpenAI!s new language generator GPT-3 is shockingly good
—and completely mindless. MIT Technology Review. https://www.technologyreview.com/
2020/07/20/1005454/openai-machine-learning-language-generator-gpt-3-nlp/
• Hoang, A. T. (2013). Impact of humor in advertising on consumer purchase decision.
• Hodges, A. (2014). Alan Turing: the enigma. In Alan Turing: The Enigma. Princeton University
Press.
• Holt, D. B. (2004). How brands become icons: The principles of cultural branding. Harvard
Business Press.

145
• Holzinger, A., Biemann, C., Pattichis, C. S., & Kell, D. B. (2017). What do we need to build
explainable AI systems for the medical domain?. arXiv preprint arXiv:1712.09923
• Hopkins, C. C. (1968). Scientific advertising. New Line Publishing.
• How We Use AI to Enhance Your Writing | Grammarly Spotlight. (2019, May 17). Grammarly
Spotlight: How We Use AI to Enhance Your Writing. https://www.grammarly.com/blog/how-
grammarly-uses-ai/
• Huang, M. H., & Rust, R. T. (2021). A strategic framework for artificial intelligence in marketing.
Journal of the Academy of Marketing Science, 49, 30-50.
• Huh, J., & Malthouse, E. C. (2020). Advancing computational advertising: Conceptualization of
the Field and future directions. Journal of Advertising, 49(4), 367–376.
• Ilkou, E., & Koutraki, M. (2020). Symbolic Vs Sub-symbolic AI Methods: Friends or Enemies?.
In CIKM (Workshops).
• Ippolito, D., Yuan, A., Coenen, A., & Burnam, S. (2022). Creative Writing with an AI-Powered
Writing Assistant: Perspectives from Professional Writers. arXiv (Cornell University). https://
doi.org/10.48550/arxiv.2211.05030
• James, W. M. (1878). Brute and human intellect. The Journal of Speculative Philosophy, 12(3),
236-276.
• Jasso, L. R. (2013). The use of archetypes in advertising: How brands can remain relevant in a
rapidly changing advertising industry through the concept of archetypes.
• Jesslyn, & Agustiningsih, G. (2021). APPLICATION OF COPYWRITING ELEMENTS IN
SOCIAL MEDIA ADVERTISING DRINKING PRODUCTS NOW IN CREATING
CONSUMER INTEREST. Jurnal Komunikasi Dan Bisnis, 9(1), 55–67. https://doi.org/10.46806/
jkb.v9i1.678
• Jones, N. (2020). my monstituces composer: Looking for the Pre-Emergent Social Consciousness
of AI in Small Data Literary Synthesis. NMC Media-n. https://doi.org/10.21900/
j.median.v16i1.222
• Joshi, N. (2019, June 20). 7 Types Of Artificial Intelligence. Forbes. https://www.forbes.com/
sites/cognitiveworld/2019/06/19/7-types-of-artificial-intelligence/The revolutionary potential of
artificial intelligence (A.I.) as it develops and becomes more integrated into our daily electronic
interactions is highlighted by Du and Xie (2021). It is vital in many industries, including

146
 healthcare, banking, and transportation, thanks to its versatility and learning potential. The authors
are nevertheless upbeat about the future of artificial intelligence despite the call for "explainable
AI" and the requirement for openness in important industries. They note that despite current
difficulties, its technological underpinnings, like computer vision and machine learning,
contribute to its continued innovation and potential to change technology.
• Kapferer, J. N. (2008). The new strategic brand management: Creating and sustaining brand
equity long term. Kogan Page Publishers.
• Kapferer, J.-N. (2012). The new strategic brand management: Advanced insights and strategic
thinking (5th ed.). Kogan Page.
• Kartsivadze, T. (2022). COPYWRITING IN SOCIAL MEDIA. International Journal of
Innovative Technologies in Economy, 4(40). https://doi.org/10.31435/rsglobal_ijite/
30122022/7924
• Katz, D. M., Bommarito, M. J., Gao, S., & Arredondo, P. (2023). Gpt-4 passes the bar
exam. Available at SSRN 4389233.
• Keller, K. L. (1993). Conceptualizing, measuring, and managing customer-based brand equity.
Journal of Marketing, 57(1), 1-22.
• Keller, K. L. and Richey, K. (2006), #The importance of corporate brand personality traits to a
successful 21st-century business”, Journal of Brand Management, Vol. 14 No. 1/2, pp. 74-81.
• Keller, K. L., & Lehmann, D. R. (2006). Brands and branding: Research findings and future
priorities. Marketing Science, 25(6), 740-759.
• Kerikmäe, T., Hoffmann, T., & Chochia, A. (2018). Legal Technology for Law Firms:
Determining Roadmaps for innovation. Croatian International Relations Review, 24(81), 91–112.
https://doi.org/10.2478/cirr-2018-0005
• Kim, W., Ryoo, Y., Lee, S., & Lee, J. A. (2022). Chatbot Advertising As a Double-Edged Sword:
The Roles of Regulatory Focus and Privacy Concerns. Journal of Advertising, 1-19.
• Kiran, K. V. B. (2021). The Transformation Of Marketing By Ai. Shanghai Ligong Daxue
Xuebao, 23(08), 391–410. https://doi.org/10.51201/jusst/21/08419
• Kissel, P., & Büttgen, M. (2015). Using social media to communicate employer brand identity:
The impact on corporate image and employer attractiveness. Journal of Brand Management,
22(9), 755–777. https://doi.org/10.1057/bm.2015.42

147
• Klein, S., Aeschlimann, J.F., Balsiger, D.F., Converse, S.L., Court, C., Foster, M., Lao, R.,
Oakley, J., & Smith, J. (1973). Automatic Novel Writing: A Status Report.
• Korteling, J., Van De Boer-Visschedijk, G. C., Blankendaal, R., Boonekamp, R., & Eikelboom, A.
(2021). Human- versus Artificial Intelligence. Frontiers in Artificial Intelligence, 4. https://
doi.org/10.3389/frai.2021.622364
• Kose, U., & Sert, S. (2016). Intelligent content marketing with artificial intelligence.
In International Conference of Scientific Cooperation for Future (No. 837-43).
• Kumar, D. P., & Raju, K. V. (2013). The role of advertising in consumer decision making. IOSR
Journal of Business and Management, 14(4), 37-45.
• Latar, N. L. (2015). The robot journalist in the age of social physics: The end of human
journalism?. In The new world of transitioned media (pp. 65–80). Cham: Springer.
• Lee, P., Bubeck, S., & Petro, J. (2023). Benefits, limits, and risks of GPT-4 as an AI chatbot for
medicine. New England Journal of Medicine, 388(13), 1233-1239.
• Lettieri, M. (2021). L!advertising che fa ridere: quando l!ironia su Facebook influenza l!intenzione
di acquisto.
• Liu, X. (2020). De-targeting to signal quality. International Journal of Research in Marketing,
37(2), 386–404.
• Liu, Y., Han, T., Ma, S., Zhang, J., Yang, Y., Tian, J., ... & Ge, B. (2023). Summary of chatgpt/
gpt-4 research and perspective towards the future of large language models. arXiv preprint
arXiv:2304.01852.
• Loeckx, J. (2016). Blurring Boundaries in Education: Context and Impact of MOOCs. The
International Review of Research in Open and Distributed Learning, 17(3). https://doi.org/
10.19173/irrodl.v17i3.2395
• Maass, A., Castelli, L., & Arcuri, L. (2000). Measuring prejudice: Implicit versus explicit
techniques. In Social identity processes: Trends in theory and research (pp. 96–116).
• Mal!ko, A. V. (2022). Legal techniques and legal technology: the complex nature of the
interaction. Gosudarstvo I Pravo, 7, 22. https://doi.org/10.31857/s102694520020992-2
• Mangold, W., Faulds, D. (2009). Social media: The new hybrid element of the promotion mix.
Business Horizons, 52(4), 357-365.

148
• Mania, K. (2022). Legal Technology: Assessment of the legal tech industry's potential. Journal of
the Knowledge Economy, 14(2), 595–619. https://doi.org/10.1007/s13132-022-00924-z
• Marchand, R. (1985). Advertising the American dream: Making way for modernity, 1920-1940
(Vol. 53). Univ of California Press.
• Mark, M., & Pearson, C. P. (2001). The Hero and The Outlaw: Building extraordinary brands
through the power of archetypes. New York: McGraw-Hill.
• Maurya, U. K., & Mishra, P. (2012). What is a brand? A Perspective on Brand Meaning.
European Journal of Business and Management, 4(3), 122-133.
• Mayor, A. (2018). Gods and robots. In Princeton University Press eBooks. https://doi.org/
10.1515/9780691185446
• McCarthy, J. (2007). What is artificial intelligence?.
• Merlo, O., Eisingerich, A. B., Gillingwater, R., & Cao, J. J. (2022). Exploring the changing role of
brand archetypes in customer-brand relationships: Why try to be a hero when your brand can be
more?. Business Horizons.
• Mogaji, E., Olaleye, S., & Ukpabi, D. (2020). Using AI to personalise emotionally appealing
advertisement. Digital and Social Media Marketing: Emerging Applications and Theoretical
Development, 137-150.
• Moore, R., & Gillette, D. (1991). King, warrior, magician, lover: Rediscovering the archetypes of
the mature masculine. New York: HarperOne.
• Moriarty, S., Mitchell, N. D., Wells, W. D., Crawford, R., Brennan, L., & Spence-Stone, R.
(2014). Advertising: Principles and practice. Pearson Australia.
• Mosca, F., & Such, J. M. (2022). An explainable assistant for multiuser privacy. Autonomous
Agents and Multi-Agent Systems, 36(1). https://doi.org/10.1007/s10458-021-09543-5
• Nautsch, A., Jasserand, C., Kindt, E., Todisco, M., Trancoso, I., & Evans, N. (2019). The GDPR
& Speech Data: Reflections of Legal and Technology Communities, First Steps Towards a
Common Understanding. INTERSPEECH 2019. https://doi.org/10.21437/interspeech.2019-2647
• Notes from the AI frontier: Applications and value of deep learning. (2018, April 17). McKinsey
& Company. https://www.mckinsey.com/featured-insights/artificial-intelligence/notes-from-the-
ai-frontier-applications-and-value-of-deep-learning

149
• Odih, P. (2007). Advertising in modern and postmodern times. Advertising in Modern and
Postmodern Times, 1-232.
• Ogilvy, D. (1983). Ogilvy on advertising. Crown Publishers.
• Ogilvy, D. (2023). Ogilvy on advertising. Welbeck Publishing Group.
• OpenAi. (n.d.). https://openai.com. Retrieved February 1, 2023, from https://openai.com
• Pan, Y. (2016). Heading toward artificial intelligence 2.0. Engineering, 2 (4).
• Pasquale, F., & Cockfield, A. J. (2018). Beyond Instrumentalism: A Substantivist Perspective on
Law, Technology, and the Digital Persona. Mich. St. L. Rev., 821.
• Pavlik, J. V. (2023). Collaborating With ChatGPT: Considering the Implications of Generative
Artificial Intelligence for Journalism and Media Education. Journalism & Mass Communication
Educator, 10776958221149577.
• Pearson, C. S. (1989). The hero within: Six archetypes we live by. Harper & Row.
• Pearson, S. (2013). SPEAKING BODIES, SPEAKING MINDS: ANIMALS, LANGUAGE,
HISTORY. History and Theory, 52(4), 91–108. https://doi.org/10.1111/hith.10689
• Pollefeys, M., & Gool, L. V. (2002). From images to 3D models. Communications of the ACM,
45(7), 50-55.
• Rabby, F., Chimhundu, R., & Hassan, R. (2021). Artificial intelligence in digital marketing
influences consumer behaviour: a review and theoretical foundation for future research. Academy
of marketing studies journal, 25(5), 1-7.
• Racine, P. L. (2018). Promoting Knowledge: A Rationale for the Copyright Protection of
Computer-Generated Works (Doctoral dissertation).
• Ramesh, A., Pavlov, M., Goh, G., Gray, S., Voss, C., Radford, A., ... & Sutskever, I. (2021, July).
Zero-shot text-to-image generation. In International Conference on Machine Learning (pp.
8821-8831). PMLR.
• Ransbotham, S., Kiron, D., Gerbert, P., & Reeves, M. (2017). Reshaping business with artificial
intelligence: Closing the gap between ambition and action. MIT Sloan Management Review,
59(1).
• Ray, T. (2020, July 14). OpenAI!s gigantic GPT-3 hints at the limits of language models for AI.
ZDNET. https://www.zdnet.com/article/openais-gigantic-gpt-3-hints-at-the-limits-of-language-
models-for-ai/

150
• Roberts, C. Exploring Brand Personality through Archetypes. Ph.D. Thesis, East Tennessee State
University, Johnson City, TN, USA, May 2010.
• Russell, S. J., & Norvig, P. (2010). Artificial intelligence: A modern approach. Prentice Hall Press.
• Rust, R. T., Rand, W., Huang, M. H., Stephen, A. T., Brooks, G., & Chabuk, T. (2021). Real-time
brand reputation tracking using social media. Journal of Marketing, 85(4), 21-43.
• Rydell, R. J., McConnell, A. R., Mackie, D. M., & Strain, L. M. (2006). Of two minds: Forming
and changing valence-inconsistent implicit and explicit attitudes. Psychological Science, 17(11),
954-958.
• Sallam, M. (2023, March). ChatGPT utility in healthcare education, research, and practice:
systematic review on the promising perspectives and valid concerns. In Healthcare (Vol. 11, No.
6, p. 887). MDPI.
• Salmerón-Manzano, E. (2021). Legaltech and lawtech: Global perspectives, challenges, and
opportunities. Laws, 10(2), 24.
• Salmerón-Manzano, E. (2021). Legaltech and Lawtech: global perspectives, challenges, and
opportunities. Laws, 10(2), 24.
• Sampson, H. (2017). A History of Advertising-From the Earliest Times. anboco.
• Sanderson, K. (2023). GPT-4 is here: what scientists think. Nature, 615(7954), 773.
• Sarker, I. H. (2022). AI-Based modeling: techniques, applications and research issues towards
automation, intelligent and smart systems. SN Computer Science, 3(2). https://doi.org/10.1007/
s42979-022-01043-x
• Schiebinger, L., Turilli, M., & Yeo, S. K. (2021). Gender bias in artificial intelligence? A
framework for critical intervention. AI & Society, 36(1), 51-60.
• Serbanescu, A. (2022). Millennials and the Gen Z in the Era of Social Media. Social Media,
Technology, and New Generations: Digital Millennial Generation and Generation Z, 61-78.
• Severin, W. J., & Tankard, J. W. (2001). Communication Theories: Origins, methods, and uses in
the mass media. New York: Longman.
• Shahriar, S., & Hayawi, K. (2023). Let’s Have a Chat! A Conversation with ChatGPT:
Technology, Applications, and Limitations. Artificial Intelligence and Applications. https://
doi.org/10.47852/bonviewaia3202939

151
• Shapiro SC. Artificial intelligence. In: Shapiro SC. (ed) Encyclopedia of Artificial Intelligence,
vol. 1, 2nd edn. New York: Wiley, 1992
• Shaw, M. (2009). Copywriting: Successful Writing for Design, Advertising, Marketing. Hachette
UK.
• Shaw, M. (2012). Copywriting Second Edition: Successful Writing for Design, Advertising,
Marketing. Hachette UK.
• Shin, D., Kim, J. H., & Cho, S. (2021). Exploring the limitations of GPT-3 and its improvement
through combining machine learning techniques. Journal of Information Science, 47(1), 3-17.
• Silva, P., Goncalves, C. S., Godinho, C., Antunes, N., & Curado, M. (2020). Using NLP and
Machine Learning to Detect Data Privacy Violations. IEEE INFOCOM 2020 - IEEE Conference
on Computer Communications Workshops (INFOCOM WKSHPS). https://doi.org/10.1109/
infocomwkshps50562.2020.9162683
• Simkus, J. (2022). Convenience sampling: Definition, method and examples. Retrieved
Oktober, 6, 2022.
• Simonite, T. (2020, July 22). AI Text Generator GPT-3 Is Learning Our Language—Fitfully.
WIRED. https://www.wired.com/story/ai-text-generator-gpt-3-learning-language-fitfully/
• Singh, J., Flaherty, K., Sohi, R.S., Deeter-Schmelz, D., Habel, J., Le Meunier-FitzHugh, K., &
Onyemah, V. (2019). Sales profession and professionals in the age of digitization and artificial
intelligence technologies: concepts, priorities, and questions. Journal of Personal Selling & Sales
Management, 39(1), 2-22.
• Siraj, S., & Kumari, S. (2011). Archetyping the brand: strategy to connect. The IUP Journal of
Brand Management, 8(3), 47-59.
• Sloman, S. A. (1996). The empirical case for two systems of reasoning. Psychological Bulletin,
119(1), 3.
• Somalvico, M. (1987). Intelligenza artificiale. Scienza & vita nuova.
• Sparks, B. A., Perkins, H. E., & Buckley, R. (2013). Online travel reviews as persuasive
communication: The effects of content type, source, and certification logos on consumer behavior.
Tourism Management, 39, 1-9.

152
• Stone, M., Aravopoulou, E., Ekinci, Y., Evans, G., Hobbs, M., Labib, A., ... & Machtynger, L.
(2020). Artificial intelligence (AI) in strategic marketing decision-making: a research agenda. The
Bottom Line, 33(2), 183-200.
• Stratton, S. J. (2021). Population research: convenience sampling strategies. Prehospital and
disaster Medicine, 36(4), 373-374.
• Sung, Y., & Kim, J. (2010). Effects of brand personality on brand trust and brand affect.
Psychology & Marketing, 27(7), 639-661.
• Sung, Y., & Tinkham, S. (2005). Brand personality structures in the United States and Korea:
Common and culture-specific factors. Journal of Consumer Psychology, 15, 334–350.
• Szczepanski, M. (2021). Economic impacts of artificial intelligence (AI). Policy Commons.
https://policycommons.net/artifacts/1334867/economic-impacts-of-artificial-intelligence-ai/
1940719/
• Taherdoost, H. (2016). Sampling methods in research methodology; how to choose a sampling
technique for research. *How to choose a sampling technique for research (April 10, 2016)*.
• Tamkin, A., Brundage, M., Clark, J., & Ganguli, D. (2021). Understanding the capabilities,
limitations, and societal impact of large language models. arXiv preprint arXiv:2102.02503
(Doctoral dissertation). (pp. 3-27). Springer, Dordrecht. (pp. 610-623)., 28(2), 96-99.
• Tashakkori, A., & Creswell, J. W. (2007). Editorial: The new Era of Mixed Methods. Journal of
Mixed Methods Research, 1(1), 3–7. https://doi.org/10.1177/2345678906293042
• Texler, O., Futschik, D., Kučera, M., Jamriška, O., Sochorová, Š., Chai, M., ... & Sýkora, D.
(2020). Interactive video stylization using few-shot patch-based training. ACM Transactions on
Graphics (TOG), 39(4), 73-1.
• Tungate, M. (2007). Adland: a global history of advertising. Kogan Page Publishers.
• Turner, S.R. (1993). Minstrel: a computer model of creativity and storytelling.
• Van Dis, E. A., Bollen, J., Zuidema, W., van Rooij, R., & Bockting, C. L. (2023). ChatGPT: five
priorities for research. Nature, 614(7947), 224-226.
• Van Esch, P., & Black, J. S. (2021). Artificial Intelligence (AI): revolutionizing digital
marketing. Australasian Marketing Journal (Amj), 29(3), 199–203. https://doi.org/
10.1177/18393349211037684.

153
• Vehovar, V., Toepoel, V., & Steinmetz, S. (2016). Non-probability sampling (Vol. 1, pp. 329-345).
The Sage handbook of survey methods.
• Wahle, J. P., Ruas, T., Mohammad, S. M., Meuschke, N., & Gipp, B. (2023). Ai usage cards:
Responsibly reporting ai-generated content. arXiv preprint arXiv:2303.03886.
• Wang, W., & Siau, K. (2019). Artificial intelligence, machine learning, automation, robotics,
future of work and future of humanity. Journal of Database Management, 30(1), 61–79. https://
doi.org/10.4018/jdm.2019010104
• Webb, J. (2020). Legal Technology: The Great Disruption? SSRN Electronic Journal. https://
doi.org/10.2139/ssrn.3664476
• Webb, J. (2020). Legal Technology: The Great Disruption?. U of Melbourne Legal Studies
Research Paper, (897).
Aaker, D. A. (1996). Building strong brands. New York: Free Press.
• Wijaya, B. S. (2012). The development of hierarchy of effects model in advertising. International
Research Journal of Business Studies.
• Woodside, A. G., Sood, S., & Miller, K. E. (2008). When consumers and brands talk: Storytelling
theory and research in. Psychology & Marketing, 25(2), 97–145. https://doi.org/10.1002/
mar.20203
• World Economic Forum. (2018). The future of jobs report. Retrieved from http://
www3.weforum.org/docs/WEF_Future_of_Jobs_2018.pdf
• Wu, J., Gan, W., Chen, Z., Wan, S., & Lin, H. (2023). Ai-generated content (aigc): A survey. arXiv
preprint arXiv:2304.06632.
• Zehnder, S. M., & Calvert, S. L. (2004). Between the hero and the shadow: Developmental
differences in adolescents!"perceptions and understanding of mythic themes in superhero comics.
Journal of Youth and Adolescence, 33(2), 105-115.
• Zeng, Y., Zhao, D., Zhao, F., Shen, G., Dong, Y., Lu, E., Zhang, Q., Sun, Y., Liang, Q., Zhao, Y.,
Zhao, Z., Fang, H., Wang, Y., Li, Y., Liu, X., Du, C., Kong, Q., Ruan, Z., & Bi, W. (2023).
BrainCog: A spiking neural network based, brain-inspired cognitive intelligence engine for brain-
inspired AI and brain simulation. Patterns, 4(8), 100789. https://doi.org/10.1016/
j.patter.2023.100789

154
• Zhai, X., Chu, X., Chai, C. S., Jong, M. S., Starčič, A. I., Spector, M., Liu, J., Jing, Y., & Li, Y.
(2021). A Review of Artificial Intelligence (AI) in Education from 2010 to 2020. Complexity,
2021, 1–18. https://doi.org/10.1155/2021/8812542
• Zhang, C., Zhang, C., Zheng, S., Qiao, Y., Li, C., Zhang, M., Dam, S. K., Thwal, C. M., Tun, Y.
L., Huy, L. L., Kim, D., Bae, S., Lee, L., Yang, Y., Shen, H. T., Kweon, I. S., & Hong, C. S.
(2023). A Complete Survey on Generative AI (AIGC): Is ChatGPT from GPT-4 to GPT-5 All You
Need? arXiv (Cornell University). https://doi.org/10.48550/arxiv.2303.11717
• Zhao, G., Li, Y., & Xu, Q. (2022). From Emotion AI to Cognitive AI. International Journal of
Network Dynamics and Intelligence, 65–72. https://doi.org/10.53941/ijndi0101006
• Zhou, L., Gao, J., Li, D., & Shum, H. (2020). The Design and Implementation of XiaoIce, an
Empathetic Social Chatbot. Computational Linguistics, 46(1), 53–93. https://doi.org/10.1162/
coli_a_00368
• Zia, A., Javed, N., & Bilal, M. (2018). Copywriting Elements and Brand Relationship: An
Analysis of Print Advertisements!"Language. Global Social Sciences Review. https://doi.org/
10.31703/gssr.2018(iii-iii).23

155