AI Social Companionship
Uploaded by
OttovonChesterfieldAI Social Companionship
Uploaded by
OttovonChesterfieldTechnological Forecasting & Social Change 193 (2023) 122634
Contents lists available at ScienceDirect
Technological Forecasting & Social Change
journal homepage: www.elsevier.com/locate/techfore
Social companionship with articial intelligence: Recent trends and
uture avenues
Rijul Chaturvedi a, Sanjeev Verma a, Ronnie Das b, Yogesh K. Dwivedi c, d, *
a
National Institute o Industrial Engineering, Mumbai, India
b
Department o Marketing, Audencia Business School, 8 Rte de la Jonelière, 44300 Nantes, France
c
Digital Futures or Sustainable Business & Society Research Group, School o Management, Swansea University, Bay Campus, Fabian Bay, Swansea, UK
d
Department o Management, Symbiosis Institute o Business Management, Pune & Symbiosis International, Deemed University, Pune, Maharashtra, India
A R T I C L E I N F O A B S T R A C T
Keywords: The social companionship (SC) eature in conversational agents (CAs) enables the emotional bond and consumer
Aective computing relationships. The heightened interest in SC with CAs led to exponential growth in publications scattered across
Social presence disciplines with ragmented ndings, thus limiting holistic understanding o the domain and warrants a
Social companionship
macroscopic view o the domain to guide uture research directions. The present study lls the research void by
Conversational agents
Anthropomorphism
oering a comprehensive literature review entailing science perormance and intellectual structure mapping. The
Articial intelligence comprehensive review revealed the research domain's major theories, constructs, and thematic structure. The-
matic and content analysis o intellectual structure resulted in a conceptual ramework encompassing anteced-
ents, mediators, moderators, and consequences o SC with CAs. The study discusses uture research directions
guiding practitioners and academicians in designing ecient and ethical AI companions.
1. Introduction agents now oer emotional support (Provoost et al., 2017) and establish
human-relational bonds with the users (Darcy et al., 2021). Thereore,
Due to digital advancements, customers now reach companies CAs oering social companionship to their users are reerred to as AI
regardless o geographical location, time, and channel does not limit the companions.
interactional continuity between rms and consumers (Suwono and The deployment o empathetic chatbots transcends dierent in-
Sihombing, 2016). Additionally, deploying Articial Intelligence (AI) dustries like banking, health care, e-commerce, education, and tourism
lits consumers' services (Ameen et al., 2021) to the next level through (Adam et al., 2021; Lee et al., 2020b; Rhee and Choi, 2020; Hsieh, 2011;
personalization and convenience. Thereore, rms deploy emerging Bickmore et al., 2013) due to its multiaceted role. In recent years, many
technologies powered by AI to oer seamless experiences to their cus- researchers have reviewed the literature on conversational agents (Lim
tomers. AI-enabled conversational agents (CAs) that provide digital et al., 2022; Rapp et al., 2021) or AI agents or health support (Gasteiger
assistance and build customer relations become signicant in such a et al., 2021). However, the literature on social companionship with
scenario. They are now evolving with new capabilities enabling them to conversational agents is scattered, thus limiting our understanding to-
engage users or prolonged interactions. wards the eld.
CAs unold paradigm shits in human-computer interaction (Biundo But why conducting a systematic literature review on companion-
et al., 2016). Leading organizations like Amazon and Google use digital ship with conversational agents is o urgent? Because brands preer to
assistants with a sense o humour (Bothun et al., 2017). Gate box, a invest in new and emerging technologies. According to the Conversa-
hologram virtual social companion, helps users to combat loneliness and tional AI market report (2021), the global market or conversational AI
oer human-like emotional support (Hirano, 2016). “Love Plus,” a is projected to reach 18.4 billion by 2026. The evolution o CAs with
videogame by Konami, allows users to build a romantic relationship advanced capabilities allows users to shit towards new and updated
with a digital character (Lowry, 2015). AI-enabled conversational versions o chatbot applications. For example, ChatGPT, launched in
* Corresponding author at: Digital Futures or Sustainable Business & Society Research Group, School o Management, Swansea University, Bay Campus, Fabian
Bay, Swansea, UK.
E-mail addresses: rijul.chaturvedi.2020@nitie.ac.in (R. Chaturvedi), sanjeev@nitie.ac.in (S. Verma), rdas@audencia.com (R. Das), y.k.dwivedi@swansea.ac.uk
(Y.K. Dwivedi).
https://doi.org/10.1016/j.techore.2023.122634
Received 1 November 2022; Received in revised orm 6 April 2023; Accepted 7 May 2023
Available online 20 May 2023
0040-1625/© 2023 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
November 2022 by Open.AI, can now remember earlier user conversa- resources, strategy, and operations (Sridevi and Suganthi, 2022; Kar and
tions. ChatGPT can write essays, poems, news articles, and stories or Kushwaha, 2021). Marketing is no exception, as chatbots advancements
users, and thus has reached one million users in just ve days o launch. and their capabilities have led to their adoption as customer service
Extant literature discusses CAs with varied unctions and capabilities. agents in the past decades. Firms deploy modern chatbots to enrich the
The question is whether investigating these capabilities can lead to a customer experience with new technologies (Kushwaha et al., 2021;
long-term relational bond, such as companionship between users and AI. Chaturvedi and Verma, 2022). Rapp et al. (2021), conducted a literature
Ultimately, an AI system's companionship development quality could review on text-based chatbots research and ound that the modern
trigger continued usage intention. AI companions can infuence user chatbot designs are heading towards providing empathy, emotional
thoughts, decisions, behaviours, purchasing patterns etc. experience, and prolonged interactions. The eld o AI now touches
Recently, researchers have examined the eects o CA capabilities upon the theory o mind and sel-awareness o articially intelligent
(like social presence, anthropomorphic eatures, interaction style, and systems (Verma et al., 2021). Emerging themes in AI-enabled technol-
media richness) on consequences like loneliness reduction, emotional ogies include eeling AI, emotional AI, empathetic AI and aective
connection, attitude towards the product, intention to use, and customer computing (Huang et al., 2019). Thereore, the uture or brands lies in
satisaction (Jones et al., 2021; Araujo, 2018; Adam et al., 2021; Rhee creating, communicating and delivering AI companions that can provide
and Choi, 2020; Lee et al., 2020a; Cheng and Jiang, 2020). Advance CAs long time emotional and unctional support to the consumers.
can now reduce the user's loneliness (Skjuve et al., 2021), provide social According to Lim (2012), AI companions are “robots or virtual
interaction support to children and people with special needs (Ramadan conversational agents that possess a certain level o intelligence and auton-
et al., 2021; Sa et al., 2021), reduce patients' loneliness (Loveys et al., omy as well as social skills, allowing them to establish and maintain long-term
2019), and provide lie support and companionship to older adults relationships with users.” The eld o human-computer interaction has
(Tsiourti et al., 2016). evolved rom interace design to social acceptability and believability
Despite available examples o companion chatbots like Replika, (Pesty and Duhaut, 2011). AI companions mitigate the loneliness o
Mitsuku, XiaoIce and Gate box, the ethical implications o such individuals seeking emotional and social support, which became sig-
advanced chatbots that can orm long-term connections with users have nicant in the Covid-19 isolation phase (Odekerken-Schröder et al.,
been the subject o ongoing debate among scholars (Murtarelli et al., 2020).
2021). While some argue that these chatbots can provide valuable social
support and companionship (Ta et al., 2020), others express concern
2.1. Conversational agents (CAs)
about the potential or manipulation and exploitation (Possati, 2022). In
such a case, an absence o a systematic literature review makes it
CAs are sotware agents designed to mimic human conversations via
challenging to address these issues comprehensively. Notably, the
natural language processing through communication channels like
consolidation o the area can unurl immense opportunities or mar-
speech, text, gestures, and acial expressions (Laranjo et al., 2018), oten
keters, practitioners, academicians, and customers. Thereore,
appear as text-based chatbots, digital avatars, and social robots (Rad-
converging literature on CAs or social companionship becomes
ziwill and Benton, 2017). CAs can be digital assistants, recommendation
essential.
agents, and social companions (McGoldrick et al., 2008). Chatbots oer
As technology is heading towards eeling AI, research on AI com-
unctional utilities that help customers perorm digital tasks like setting
panions prolierates. However, the literature still lacks a comprehensive
the alarm or reminder, checking the weather, playing songs, searching
ramework to unveil SC's antecedents, mediators, moderators, and out-
or inormation, including product recommendations, etc. (Chaturvedi
comes with CAs. Moreover, current trends and ways orwards remain
and Verma, 2022). CAs assist consumers in ordering products via voice
uzzy without a systematic literature review. The ollowing research
commands (Aw et al., 2022). Aective computing advancements
questions remain unanswered:
enabled CA to build emotional connections with humans. According to
RQ1. . What are the publication and citation trends in SC with CAs? Hamilton et al. (2021), humans preer recommendations rom known
ties (riends or companions in their network) and repeat purchase
RQ2. . Which are the top sources, publications, and authors in SC with
intention based on voice assistants recommendations leads to brand
CAs?
loyalty (Maroukhani et al., 2022). Thus, advanced AI companions can
RQ3. . What are the major theories in SC with CAs? maintain long-term relationships with humans, moderate their emo-
tions, and induce purchase intentions and brand loyalty (McLean et al.,
RQ4. . What are the major themes in SC with CAs?
2021).
RQ5. . What are the antecedents, mediators, moderators, and outcomes o
SC with CAs?
2.2. Social companionship (SC)
RQ6. . Which directions should uture research pursue to advance the SC
with CAs? According to Benyon and Mival (2010), SC reers to “a pleasant and
accessible relationship with an interactive source, emerging out o the
The present study attempts to ll the void and discern the holistic
social and emotional investment o a person which requires a level o
view o the research domain with perormance and intellectual structure
trust, compatibility, and amiliarity with the source that results in a
mapping. The thematic and content analysis enables the bird's eye view
eeling o security, and general wellbeing.” CAs designed or SC are
o the domain (Donthu et al., 2021a).
called “Articial Companions (AI Companions),” which can substitute
The present study is organized into nine sections. Section 2 presents
human relations by observing users' past experiences (Campos and
the theoretical background, Section 3 discusses the review methodol-
Paiva, 2010). For example, in-home companions help users schedule and
ogy, Section 4 ocuses on perormance analysis, Section 5 details theo-
perorm routine health care activities, mobile companions help in out-
retical oundations, Section 6 unolds the intellectual structure, Section
door physical activities, and virtual cooking companions help by rec-
7 proposes a conceptual ramework, Section 8 discusses uture research
ommending daily recipes (Turunen et al., 2011). AI companion toys
directions, and Section 9 concludes the article.
engage children in long-term relationships (Adam et al., 2010).
“Replika: My AI riend” is an advanced AI companion with therapeutic
2. Theoretical background
resources and partners to reduce users' loneliness (Skjuve et al., 2021).
Interaction with the social chatbot (like Mitsuku) reduces the need or
AI applications are revolutionizing management streams like human
physical, social presence (Croes and Antheunis, 2021).
2
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
2.3. Evolution o AI companions support to its users (Hirano, 2016).
The ourth and most recent shit is observed with the introduction o
Over the past ew decades, AI companions have come a long way in generative AI into CAs. The recent launch o ChatGPT equipped with the
terms o their capabilities and acceptance in society. The current study, third generation o the generative pre-trained transormer model and
presents our shits in the evolution o AI companions. In 1996, the sel-supervised learning in November 2022 reached one million users in
world saw the rst AI companion in the orm o Tamagotchi (Bloch and the rst ve days. The CA can generate original content, respond to a
Lemish, 1999), a small virtual pet that users could take care o on a LED- wide range o prompts and questions, and discuss any subject as long as
based digital screen. The toy was designed to simulate the experience o the user wants (Dwivedi et al., 2023). It can also remember previous
caring or a virtual pet or which users could develop a sense o re- conversations with the user. Fig. 1 represents the our technological
sponsibility and attachment. While it used a simple design with rule- shits rom 1996 to 2022.
based modeling and physical buttons or user interaction, it sparked
an interest in having a digital companion. As technology advanced, the 3. Research methodology
capabilities, architecture and companionship eatures improved.
Table 1, summarizes the evolution o AI companions rom 1996 to 2022. Systematic literature review (SLR) is an appropriate approach to
In 2001, Baby harp Seal PARO, a therapeutic robot, entertained address the research questions pertaining to current trends and uture
patients in hospitals, older adults and other lonely individuals research directions in the research domain (Whittemore et al., 2014).
(Takayanagi et al., 2014). The robot is designed to look like a baby Seal The present study delves into the literature on SC with CAs to uncover
with real ur. Its architecture includes uzzy logic and reinorcement the perormance analysis (authors, sources, and documents) and intel-
learning techniques, enabling PARO to behave like a real seal. Although lectual structure (themes, constructs, and theories) o the current liter-
the design o the robot is limited in the range o actions and expressions, ature (Donthu et al., 2021a). The scientic approach makes the review
it provides signicant emotional support to the users and allows them to replicable, transparent, and objective. The present study replicable
love and care or someone. Nabaztag, released in 2008, is a rabbit- research protocol contains ve steps o in-depth analysis and adheres to
shaped smart device that can announce weather orecasts and news established guidelines in previous studies (Verma and Yadav, 2021;
headlines, play MP3 streams, and send and receive messages (Cavazza Verma et al., 2021; Mhatre et al., 2020) to reveal current trends and the
et al., 2008). User interaction with a rabbit inormation console pro- way orward. Fig. 2 highlights the fow process to conduct the current
vided a sense o care or the rabbit and ormed a bond with it over time. literature review.
The technological evolution o AI companions then shited to
anthropomorphic characters with the release o KASPAR, a child-shaped 3.1. Stage 1: search strategy
doll-like robot in 2005 (Wood et al., 2019). The robot is specially
designed or Kids with autism disorder that allow users to love and care The search strategy delimits source types to only journals to retrieve
or the robot with responsive acial expressions on the human touch. scientic, contemporary, and explanatory literature (Chandra et al.,
Such companions possess limited conversational capabilities, thus 2022; Lim et al., 2022). For literature search and retrieval, the biblio-
restricting users rom enhanced interactions. With natural language metric database Scopus, instead o other alternatives, such as the web o
processing (NLP) and deep learning-based modeling techniques, Apple science, google scholar, etc., is preerred, as Scopus ensures the stringent
released Siri with a technological breakthrough in 2011 (Thorne, 2020). criteria or indexation o published documents (Verma, 2022; Chandra
Siri assists users with digital tasks such as making phone calls, setting et al., 2022; Lim et al., 2022).
alarms, playing music and videos etc. The virtual assistant was only
available or Apple devices, but its unctional utility helped users 3.2. Stage 2: selection o search string
manage their digital tasks. Amazon launched Alexa in 2014 with
recurrent neural networks (RNNs) and learning-based modeling to A pool o keywords refecting the research domain and the possible
provide a richer user experience to its customers (Gao et al., 2018). The synonyms o keywords ormed the search string. The search string was
unctional intelligence o Alexa and its humorous responses to the users also added to the keyword's repository used in similar review studies in
enabled it to improve proximity with amazon customers. Despite a extant literature. Academic experts were consulted or the nalization o
complex design with many unctions, such as intelligent home auto- the search string. The search was ocused on the social companionship
mation and digital assistance, Alexa has yet to understand user emotions aspects o the conversational agents and thus keywords such as chatbot
and respond accordingly. or social bot were excluded rom the search to avoid a large number o
The third shit in the evolution o AI companions observed the irrelevant results. Search strings include keywords like “Articial
introduction o aective computing into chatbot designs. Microsot Asia Intelligent agents” OR “AI agents” OR “Digital Assistants” OR “Virtual
released an empathetic chatbot called XiaoIce in 2014. The CA is Assistants” OR “Conversational AI” OR “Conversational agents” OR “AI
designed to provide emotional support to the users with its ability to companion” OR “Digital Companion” OR “Virtual Companion” OR
understand user emotions through sentiment analysis (Zhou et al., “Articial Companion” AND “Companion” OR “Love” OR “Friend” OR
2020). Modern architecture with Convolutional neural networks (CNNs) “Bond” OR “Emotion”. In May 2022, the search perormed in the Scopus
and long-short-term memory networks (LSTM) allowed XiaoIce to Bibliometric database resulted in 988 documents. The resultant litera-
engage users or prolonged interaction and develop human-like bonds. ture database was downloaded in Bibtex (.bib) and comma-separated
However, the chatbot only continues the discussion or the user until value (.csv) le ormats or urther processing.
input rom the user is received. Replika is an advanced version o the
emotional chatbot launched in 2016. Its unique design allows AI to 3.3. Stage 3: ltering the initial results (inclusion and exclusion criteria)
mimic human romantic conversations with the users (Ta et al., 2020).
One technological breakthrough with Replika is its ability to initiate The Scopus database unctions, such as sorting and ltering, acili-
conversations with the user without input, just like humans. RNNs, tated the organization o codes or language, subject area, document
LSTM and reinorcement learning techniques allow Replika to provide type, and source type. Two-stage ltering instrumentalizes the precise
its users with emotional and mental health support up to a level that can data collection. Initially, ltering is operationalized on the ltering
reduce eelings o loneliness. With one o the most advanced chatbot menu available in the Scopus database, ollowed by manual ltering
designs, Replika has yet to learn about digital assistance and home through careul scanning o each document. In the initial ltering,
automation skills. Gatebox, a holographic virtual assistant launched in documents were delimited to publications in the English language;
2016, has both the capabilities that provide utilitarian and emotional document type as article or review; source type as journals and
3
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Table 1
Evolution o AI companions rom 1996 to 2022.
Characters Articial pets Anthropomorphic characters
Tamagotchi (Bloch and PARO (Takayanagi et al., Nabaztag (Cavazza et al., KASPAR (Wood et al.,
Name Apple's Siri (Thorne, 2020)
Lemish, 1999) 2014) 2008) 2019)
Time 1996 2001 2005 2005 2011
Theme Virtual pet Therapeutic robot Inormation console Therapeutic robot Digital assistant
A virtual Pet in a small egg- A baby harp seal robot with A rabbit-shaped smart
A child-shaped doll- A virtual assistant in apple smart
Description shaped toy with a mini led real ur, fippers, and device with two long ears
like robot. phones
screen. vocalizations. that could move.
Fuzzy logic and reinorcement
Architecture Rule-based modeling Rule-based modeling Rule-based modeling Deep learning-based modeling
learning
The weather orecast, stock Voice command, inormation
market report, news search, setting alarms,
The user could eed the pet, Singing songs,
Recognizing human voices, headlines, alarm clock, e- reminders, making phone calls,
Capabilities play games with it, and even recognizing the human
human touch, mail alerts, sending and announcing weather orecasts,
discipline it i it misbehaved. touch
receiving messages, and playing music and videos, and
MP3-Streams. third-party app integration.
Appearance Virtually embodied Physically embodied Physically embodied Psychically embodied Disembodied
Medium o Voice-based, LED lights
Physical buttons on the toy Touch Sensors Touch Sensors Text, Image, and Voice
interaction display, and web interace
A portable toy with a
Robot ocuses on emotional Integration o Wi- Conversational ability to
microprocessor could allow
Key specialty support by mimicking the connectivity into a Expressive ace understand the natural language
users to interact with a virtual
behaviour o a real baby Seal. consumer device o users and reply accordingly
pet.
The toy was designed to
Siri's ability to learn about users
simulate the experience o Baby harp seal that users An articial rabbit that Human child-like
over time and its integration
Companionship caring or a virtual pet or could love and care or by users could interact with, appearance with acial
with various third-party apps
eature which users could develop a moving their hands on its ur care or and orm bond expressions that Kids
and services made it a more
sense o responsibility and and receiving its reactions. with over time could love and care or.
helpul assistant or users.
attachment.
The device could not Siri is unable to remember
Small monochrome screen A limited range o actions and Limited interaction
display videos, images, and earlier conversations with users.
with limited graphics. Basic expressions also required abilities as it was
Key limitations text, also limited in Lack o sel-disclosure and
medium o interaction with the regular maintenance and designed or kids with
interacting with the users emotional engagement. It is
help o buttons only. charging. autism disorder.
aectively. limited to Apple devices.
Characters Anthropomorphic characters AI characters
Amazon's Alexa (Gao et al., Microsot's XiaoIce (Zhou Gate box (Hirano, ChatGPT
Name Replika (Ta et al., 2020)
2018) et al., 2020) 2016) (Dwivedi et al., 2023)
Time 2014 2014 2016 2016 2022
Theme Digital assistant Empathetic chatbot Personal companion Personal companion Versatile chatbot
A versatile AI chatbot that can
write poems, stories, news
A virtual assistant in smart A companion chatbot that A holographic
A riendly chatbot who can articles and generate original
Description home devices and smart mimics romantic domestic companion
understand user emotions. text on any subject. It can also
phones. conversations and smart assistant
play various roles like travel
advisor, book author act.
Learning based
Learning Based Modeling, Convolutional neural
modeling, computer Generative Pre-trained
Recurrent Neural Networks networks (CNNs), long-short LSTM, RNNs, and
Architecture vision, robotics, Transormer, sel-supervised
(RNNs), Long-short term term memory networks Reinorcement learning.
internet o things, learning
memory networks (LSTM) (LSTM)
emotion recognition
Functional intelligence to Capable o generating original
Emotional intelligence, Emotional intelligence,
perorm tasks such as Named entity recognitions, content, can respond to wide
mental health support and Functional intelligence
switching o or on lights and sentiment analysis and range o prompts and questions,
gamication allows in home automation,
Capabilities other appliances, playing emotional intelligence enable can discuss any subject as long
Replika to mimic romantic mental health support,
music, making phone calls, XiaoIce to engage users or as user wants. Can remember
conversations with the social media
announcing weather, setting prolonged conversations. previous conversations with the
users. integration
alarms etc.. user.
Holographic
Appearance Disembodied Virtually embodied Virtually embodied Disembodied
embodiment
Medium o
Text, voice, image Text, voice, image Text, voice, image Text, voice, image Text
interaction
Focus on emotional
A physical appearance
Alexa skills kit to integrate To generate emotionally intelligence, empathy, sel-
o avorite character Generative AI is the key
Key specialty Alexa with wide range o smart responsive and contextually disclosure and mimicking
that users can live specialty in ChatGPT
home devices. relevant responses human romantic
with.
conversations.
Emotional support, mental Emotional support, Providing writing support,
Voice based device controls, Emotional support and ability
health support and ability mental health support, inormation support and the
and humorous responses o to engage users in prolonged
Companionship to converse romantically and home automation ability to play versatile roles
Alexa make user-Alexa interactions enable XiaoIce to
eature allows Replika to make support allow users to enables ChatGPT to establish
relationship proximal in orm relational bonds with the
users eel emotionally consider AI a living long-term relationships with the
nature. users
connected. partner. users.
(continued on next page)
4
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Table 1 (continued )
Characters Articial pets Anthropomorphic characters
XiaoIce is only limited to
Lack o contextual Japanese and Chinese
Lack o unctional Only limited to
understanding, cannot answer languages. Responses Lack o unctional abilities,
intelligence like Alexa, Japanese and English
Key limitations long and complex queries and sometimes eel scripted or disembodiment and limited to
limited conversational skill language, relatively
lack o emotional repetitive. Lack o unctional text based interaction medium.
to stretch a discussion. expensive to buy.
understanding. intelligence like Alexa. Lack o
sel-disclosure.
employs network analysis to understand the interrelationship between
documents, keywords, authors, or citations (Pilkington and Catherine,
1999). The data clustering algorithm is the oundation o network
analysis to discern the intellectual structure o the research domain
(Chen et al., 2010). Nodes in a cluster possess similar characteristics in
the network (Radicchi et al., 2004). The network parameters, such as
edges' thickness and density, depict the similarity index between the
scientic actors (Leydesdor and Raols, 2011; Radicchi et al., 2004).
Betweenness centrality identies the most prominent nodes in the
shared linkage (Hjørland, 2013). Network's modularity index is calcu-
lated using the Louvain algorithm. Leading eigenvalues, or spin glass
algorithms, measure the strength o the relationship between nodes
(Blondel et al., 2008). We preer the Louvain algorithm in the present
study as it optimizes the network run in time 0 (n log n). The sotware
eliminates the outliers o the network present in the orm o isolated
nodes on the map. The algorithm returns an optimized value with the
network's modularity ranging rom 1 to +1. For a weighted graph, the
modularity is ormulated as
( )
Fig. 1. Technological shits in the evolution o AI Companions. 1 ∑∞ ki kj
Q= aij ᵟ(CiCj)
2m ij 2m
conerence proceedings; and subject area as “Social science, Psychology,
Decision Science, Arts and Humanities, Neurosciences, Health Pro- 4. Performance analysis of SC with CAs
essions, Business, Management and Accounting, Economics Econo-
metrics and Finance, and Multidisciplinary.” Manual ltering involves Most reviews examine the perormance o research constituents, e.g.,
careul scanning o documents or the relevancy and specicity o the authors, journals, and documents (Lim et al., 2022; Donthu et al., 2021a;
research domain. Two-stage ltering led to a reduced number o 197 Chandra et al., 2022), to show the signicance o the chosen eld. The
documents. current study attempts to unurl publication and citation trends or
research on SC with CAs. According to Mukherjee et al. (2022), top
3.4. Stage 4: data collection, data cleaning, and data processing publications and top authors o the research eld support practitioners
in identiying signicant contributions in the given eld, along with a
Post-screening, the data set was triangulated or data verication. list o experts in that domain. Additionally, a list o top sources helps
Two researchers and a group o experts opined the suitability o data. researchers and academicians to target suitable journals or publication
Finally, 126 papers ormed the consideration set or urther processing. o their studies in the given eld. In this regard, the current study maps
The present study uses an inductive approach to drive data insights (Far the perormance analysis o SC with CAs in the ollowing sub-sections.
and Rad, 2018; Lim et al., 2022). Besides the inductive approach,
deductive analysis helped in perormance analysis and science mapping 4.1. Publication and citation trend o SC with CAs (RQ1)
o the research domain (Donthu et al., 2021a). Perormance analysis
reveals the publication trend by identiying the domain's top authors, Table 2 presents the publication and citation trends or research on
documents, and sources. Content and thematic analysis revealed the SC with CAs. The table indicates that the eld is about nineteen years
prevailing theories, constructs, keyword co-occurrence, and biblio- old. The initial papers started appearing in 2003, and the domain
graphic coupling themes. The present study uses Microsot excel or evolved gradually. The dataset contains 126 documents (TP) published
content analysis, Biblioshiny R sotware, and VOSviewer sotware or a in 87 dierent sources, with 106 cited publications (TCP) and a pro-
network o themes (Donthu et al., 2021a). ductivity average o 6.63 publications per year. Fig. 3 represents the
publication trend and indicates the sharp growth in the last three years.
3.5. Stage 5: data analysis strategy The exponential growth in the last three years indicates the growing
interest in the social companionship aspect o conversational agents.
The study relies on the keyword Co-occurrence network and biblio- The involvement o digital assistants in consumers' lives can be attrib-
graphic coupling analysis to reveal dierent knowledge clusters present uted to social abrics and emotional connections developed by conver-
in the domain. Network analysis is perormed through VOS viewer, sational agents.
which creates clusters o keywords or documents based on their weight The table indicates that studies on SC with CAs have received 2231
(Van Eck and Waltman, 2011) and supports the content analysis. The citations (TC), with an average citation per publication o 17.70 (TC/
bibliographic coupling is a widely preerred technique that reveals the TP). The h-index (citation impact) o the eld 25 (h) inorms that
thematic analysis (Boyack and Klavans, 2010). The present study twenty-ve publications have received at least 25 (h) citations. In
5
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Fig. 2. Highlights the fow process or the review.
contrast, the eld's g-index (citation infuence) is eight, which indicates in 2018). The research papers published in Computers in Human
that eight publications have received at least 64 (g2) citations each. The Behaviour garnered the highest average number o citations (C/Y:
authorship inormation in the table reveals a total o 428 authors 57.20) per year, signiying the journal as the leading source or publi-
(including repetition) (NCA) or 391 unique authors (excluding repeti- cation in the given eld. Social companionship entails behavioural as-
tion) (NUA) contributing to the eld. The dataset includes teen single- pects with AI-enabled conversational agents rationalizing the higher
authored publications (SA), while 111 documents are co-authored publications in journals/conerences ocusing on the interace between
publications (CA). The collaboration index (CI=NCA-TP÷TP) o 3.14 computers and human behaviour. Conerence on human actors in
signies that each lead author has collaborated with an average o 3.14 computing systems proceedings is the most impactul conerence source
co-authors (CI). Most o the documents in the dataset are empirical (h-index = 5 and g-index = 5) or researchers working on SC with CA.
studies (92.85 %), while only 9 (7.14 %) are non-empirical. Noteworthily, the top 10 sources or publications on social compan-
ionship with conversational agents include the majority o sources (40
%) rom the area o Computer Science (CS) (i.e., Computers in Human
4.2. Top sources or SC with CAs (RQ2) Behaviour, International Journal o Human-Computer Studies, Conerence
on Human Factors in Computing Systems – Proceedings and Conerence on
Table 3 presents the top ten sources or publications in SC with CAs Human Factors in Computing Systems – Proceedings and Lecture Notes in
based on the highest number o citations received. The table indicates Computer Science) ollowed by the sources rom the area o inormation
that Computers in Human Behaviour receives the highest number o systems (IS) (i.e., Frontiers o Inormation Technology and Electronic
citations (TC: 286), with only ve publications in a short period (starting
6
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Table 2 Thereore, the current study goes beyond the number o citations and
Bibliometric inormation o SC with CAs. checks other research quality indexes such as solidity and plausibility,
Panel A. Publication inormation Statistic originality and novelty, scientic value, and societal value and rele-
vance (Aksnes et al., 2019). The table indicates Araujo's (2018) exper-
Total publications (TP) 126
Total cited publications (TCP) 106 imental study with chatbots is the most impactul publication with the
Total sources (TS) 87 highest number o citations (222) in the eld. The study is also the most
Number o active years (NAY) 19 1/2 infuential publication, with the highest average citations per year o
Productivity per active year (PAY) 6.63 44.40, which ound that social presence mediates the eect o anthro-
pomorphic design cues on the emotional connection established with the
Panel B. Citation inormation Results users (Araujo, 2018). For solidity and plausibility, we checked the
Total citations (TC) 2231
quality o citing journals. We ound that they are journals o repute like
Average citations per publication (TC/TP) 17.70
h index 25 the Journal o the Academy o Marketing Science, Journal o Management
g index 44 Inormation Systems, International Journal o Human Resource Manage-
ment, Electronic Markets, Journal o Business Research, Journal o Retailing
Panel C. Authorship inormation and Consumer Services, Psychology & Marketing, and Journal o Service
Number o contributing authors (including repetition) (NCA) 428 Management. In terms o originality and novelty, the study emphasizes
Number o unique authors (excluding repetition) (NUA) 391 the novel concept o anthropomorphic design in chatbots and its sig-
Authors o single-authored publications (ASA) 13
nicance in business, which is in triangulation with the evolution o AI
Authors o co-authored articles (ACA) 378
Single-authored publications (SA) 15
companions explained in the prior sections o the current paper.
Co-authored publications (CA) 111 Regarding scientic value, we conrmed that the authors citing the
Collaboration index (CI=NCA-TP÷TP) 3.14 study are established academicians and researchers. We used alternative
metrics through “PlumX” to assess the societal value and relevance o
Panel D. Document inormation the article and ound that the study captured the attention o 889 readers
Article (empirical) 117
Reviews (non-empirical) 9
Reerences 7365
Table 3
Keywords 420
Top sources or SC with CAs.
Note: Period o coverage = 2003 – May 2022.
Journals TC h g TP Start_PY C/Y
Computers in Human Behaviour 286 4 5 5 2018 57.20
Engineering, Computer Networks, Computational Linguistics, European
Frontiers o Inormation Technology
Conerence on Inormation Systems). While CS and IS are traditionally and Electronic Engineering 206 1 1 1 2018 41.20
considered dierent elds, there is an increasing need or interdisci- International Journal o Human-
plinary research in developing and implementing AI technologies or Computer Studies 151 3 4 4 2019 37.75
Computer Networks 116 1 1 1 2013 11.60
companionship. By catering to interdisciplinary audiences rom mar-
Computational Linguistics 94 1 1 1 2020 31.33
keting, philosophy, computer science, social science and inormation Electronic Markets 64 1 1 1 2021 32
science the study aims to contribute to a more comprehensive under-
standing o the subject matter, which can have implications or the
Conerences
advanced design o CAs capable o establishing relational bonds with Conerence on Human Factors in
users in range o disciplines(Bracken and Oughton, 2006; Keestra, 2017; Computing Systems - Proceedings 137 5 5 5 2004 7.21
MacLeod, 2018). Proceedings o Conerence on Human
Inormation Interaction and
Retrieval 92 1 1 1 2018 18.40
Lecture Notes in Computer Science 63 2 7 7 2011 5.25
4.3. Top publications or SC with CAs (RQ2) European Conerence on Inormation
Systems 62 2 2 2 2018 12
The top ten publications on SC with CAs' based on the highest
Abbreviations: C/Y, citations per year; g, g-index; h, h-index; Start_PY, the start
number o citations are presented in Table 4. However, citation counts
o publication year; TC, total citations; TP, total publications.
alone are insucient to evaluate any published work's research quality.
Fig. 3. Annual publication trend o SC with CAs' research.
7
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Table 4 article's societal value and relevance. The other top publications in
Top publications or research on SC with CAs. Table 4 deal with designing and implementing various empathetic
Title Authors Year TC C/Y chatbots and caring machines that can establish a close relationship with
humans. These include research studies by Boshma et al., 2013, Zhou
Living up to the chatbot hype: the
infuence o anthropomorphic et al., 2020, and Bickmore and Picard, 2004. Additionally, some publi-
design cues and communicative Araujo 2018 222 44.40 cations in the list highlight using social chatbots to improve customer
agency raming on conversational service (Adam et al., 2021; Bickmore et al., 2011; Lopatovska and Wil-
agent and company perceptions liams, 2018; Gnewuch et al., 2018). As mentioned or the top three, the
From Eliza to XiaoIce: challenges and
opportunities with social chatbots
Shum et al. 2018 206 41.20 authors have check the research quality indexes or each top publication
A taxonomy o social cues or in the table.
Feine et al. 2019 123 30.75
conversational agents
Design and analysis o a social botnet Boshma et al. 2013 116 11.60
The design and implementation o 4.4. Top authors or SC with CAs' research (RQ2)
Xiaoice, an empathetic social Zhou et al. 2020 94 31.33
chatbot
The top 10 authors in the eld o SC with CAs are listed in Table 5.
Personication o the Amazon Alexa: Lopatovska and
BFF or a mindless companion? Williams
2018 92 18.40 The table indicates Bickmore as the most prolic author with (4) pub-
AI-based chatbots in customer service lications and (180) total citations starting rom 2004. The list also
Adam et al. 2021 64 32.00
and their eects on user compliance contains our authors with three publications (Broadbent E, Loveys K,
Bickmore and Wilks Y, and Clavel C) and 5 authors with two publications. Li D (TC:
Towards caring machines 2004 64 3.37
Picard
Faster is not always better:
300) and Shum HY (TC: 300) are the most-cited authors in the eld,
understanding the eect o dynamic ollowed by Bickmore T (TC: 180), Gnewuch U (TC: 173), Maedche A
Gnewuch et al. 2018 50 10.00
response delays in human-chatbot (TC: 173), and Morana S (TC: 173). However, citation counts, and
interaction related traditional metrics limits the view o overall impact o authors on
Relational agents improve engagement
an emerging eld o research. Thus, the current study also investigates,
and learning in science museum Bickmore et al. 2011 50 4.17
visitors into the proles o authors and nds that, top authors in the list come
rom diversied workplaces such as academics (e.g., Bickmore T,
Note: Abbreviations: C/Y, citations per year, TC, total citations.
Broadbnet E, Loveys K, Wilks K, Clavel C, Gnewuch U, Maedche A,
Morana) and practice (e.g., Li D and Shum HY). Moreover, most o the
rom dierent parts o the globe. authors in the list are rom Western and European countries (e.g., the
Shum et al. (2018) paper is the second most impactul and infuential United States (4), New Zealand (2), Germany (3)). The current study can
publication, with (206) total citations and an average o 41.20 per year. be considered an extension o Lim et al. (2022) study on conversational
The authors presented the development o chatbots, starting rom the commerce research. The ndings align with the previous survey that
rst chatbot Eliza (1966), to the latest and most advanced social chat- western countries contribute more to research on SC with CAs.
bots like Siri (2011) and XiaoIce (2014). The authors also ound XiaoIce
as an empathetic chatbot that can recognize human emotions and
5. Theoretical foundations of SC with CAs (RQ3)
engage users or longer. The paper is cited by publications in various
journals o repute, such as the Journal o the Academy o Marketing Sci-
Table 6 lists theories used in the literature when addressing SC with
ence, Journal o Business Research, International Journal o Consumer
CAs. The table indicates that research on SC with CAs has adopted
Studies, and Knowledge-Based Systems, which ensures the solidity and
theories rom disciplines such as
plausibility o the study. In terms o originality and novelty, the study
emphasizes the novel concept o empathetic chatbots and their design
• Theories rom Psychology: Balance theory, oot-in-the-door technique,
using aective computing techniques, again in the triangulation with
uncanny valley theory, sel-perception theory, sel-disclosure theory,
the evolutions o the AI companions. The study is cited by numerous
practitioners in the eld due to the orientation o the study towards
Table 5
uturistic technology and thus ensures its scientic value. Alternative
Top authors or SC with CAs.
metrics rom PlumX indicates that the study has one patent amily, once
mentioned in a news article, tweeted two times on twitter, and has Authors Author aliations NP TC PY_start C/Y
captured 797 reading attentions throughout the globe, which ensures Northeastern University, United
Bickmore T 4 180 2004 9.47
the societal value and relevance o the article. States
Broadbent University o Auckland, New
Feine et al. (2019), with 123 total citations, is the third most-cited 3 15 2019 3.75
E Zealand
paper, with an average o 30.75 citations per year. The authors University o Auckland, New
Loveys K 3 15 2019 3.75
converged the diversied literature on social cues (such as gender, age, Zealand
gesture, etc.) that trigger humans to react while interacting with chat- Florida Institute or Human-
bots and classied them into our main categories (verbal, auditory, Wilks Y Machine Cognition, United 3 13 2005 0.72
States
visual, and invisible) and ten sub-categories. The paper is cited by Polytechnic Institute o Paris,
publications in various journals o repute, such as the International Clavel C 3 8 2013 0.8
France
Journal o Inormation Management, Decision Support Systems, Journal o Microsot Corporation, United
Li D 2 300 2018 60
Retailing and Consumer Services, Electronic markets, Journal o Manage- States
Microsot Corporation, United
ment Inormation Systems, and Psychology & Marketing etc., which ensures Shum HY 2 300 2018 60
States
the solidity and plausibility o the study. The study summarizes the Karlsruhe Institute o
Gnewuch U 2 173 2018 34.6
novel concept o social capabilities (using social cues in the chatbot Technology, Germany
design) in terms o originality and novelty. Authors rom diverse disci- Maedche A
Karlsruhe Institute o
2 173 2018 34.6
plines, including established and new, cite the study in their work, thus Technology, Germany
Karlsruhe Institute o
ensuring its scientic value. Alternative matrices rom PlumX indicates Morana S
Technology, Germany
2 173 2018 34.6
that the study has captured 452 reading attentions throughout the globe,
42 attentions on Facebook, and ve tweets on Twitter, ensuring the Abbreviations: C/Y, citations per year; Start_PY, the start o publication year; TC,
total citations.
8
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Table 6 Table 6 (continued )
Theories in SC with CAs' research. Year Field Theory Origin Example
Year Field Theory Origin Example
Dat and
Media and Media Richness Hsieh and
Psychology 1986 Lengel
Communication Theory Lee (2021)
Heider Rapp et al. (1986)
1946 Psychology Balance Theory
(1946) (2021) Walther and
Media and Social Inormation Lee et al.
Freedman 1992 Burgoon
Foot in the Door Adam et al. Communication Processing Theory (2020a)
1966 Psychology and Fraser (1992)
Technique (FITD) (2021)
(1966) Silverstone Brause and
Media and Domestication
Uncanny Valley Ta et al. 1992 and Hirsch, Blank
1970 Psychology Mori (1970) Communication Theory
Theory (2020) 1992 (2020)
Sel-Perception Adam et al. Media and Smith Porra et al.
1972 Psychology Bem (1972) 2003 Speech Act Theory
Theory (2021) Communication (2003) (2020)
Premack Interpersonal
Media and McCroskey Lei et al.
and Lee et al. 2006 Communication
1978 Psychology Theory o Mind Communication et al. (2006) (2021)
Woodru (2020a) Theory
(1978) Communication Petronio
Media and Ha et al.
Sel-Disclosure Derlaga and Lee et al. 2008 Privacy Management and Durham
1987 Psychology Communication (2021)
Theory Berg (1987) (2020c) Theory (2008)
Barrett and
The Functionalist Crolic et al.
1987 Psychology Campos
Theory o Emotions (2022) Marketing and computer science
(1987)
Cheng and
Burgoon Croes and Eisenhardt
Expectancy Violation 1989 Marketing Agency Theory Jiang
1988 Psychology and Hale Antheunis (1989)
Theory (2020)
(1988) (2021)
Brand Personality Aaker Youn and
Potdevin 1997 Marketing
Cognitive Load Sweller Theory (1997) Jin (2021)
1988 Psychology et al.
Theory (1988) Mick and Wilson-
(2021) Technology Paradox
1998 Marketing Fournier Nash et al.
Cole and Bickmore Theory
Theory o Relational (1998) (2020)
1996 Psychology Bradac et al.
Satisaction Mensio
(1996) (2011) Computer Aective Computing Lisetti
1998 et al.
Sel Determination Ryan and Sinoo et al. Science Theory (1998)
2000 Psychology (2018)
Theory Deci (2000) (2018)
Commitment- Cialdini Adam et al.
2001 Psychology
Consistency Theory (2001) (2021)
Preston and
expectancy violation theory, the unctionalist theory o emotions,
Perception-Action Gama et al. theory o relational satisaction, sel-determination theory,
2002 Psychology De Waal
Model (PAM) (2011)
(2002) commitment-consistency theory, theory o mind, perception-action
Potdevin model, emotional response theory, three-actor theory o anthropo-
Emotional Response Mottet et al.
2006 Psychology et al.
Theory (2006) morphism, and cognitive load theory.
(2021)
Three-Factor Theory Pradhan • Theories rom Sociology: Bourdieu's social theory, social penetration
Epley et al.
2007 Psychology o et al. theory, social exchange theory, relationship development theory,
(2007)
Anthropomorphism (2019) social role theory, time interaction and participation TIP theory o
groups, social response theory, computers are social actors' theory,
Sociology Aristotle's theory o riendship, and theory o companions.
Tassiello • Theories rom Media and communication: Communication privacy
Social Exchange Homans
1958 Sociology et al.
Theory (1958) management theory, interpersonal communication theory, speech
(2021)
Altman and Skjuve act theory, media richness theory, social inormation processing
Social Penetration
1973 Sociology Taylor et al. theory, domestication theory,
Theory
(1973) (2021) • Theories rom Marketing: Agency theory, brand personality theory,
ABCDE Model o Croes and technology paradox theory and computer science (e.g., aective
Levinger
1980 Sociology Relationship Antheunis
Development
(1980)
(2021)
computing theory)
Bourdieu's social Bourdieu Possati
1980 Sociology
theory (1980) (2022) The earliest theory stems rom psychology in 1946 (i.e., balance
Time Interaction and theory) and the latest theory rom sociology in 2011 (i.e., theory o
McGrath Wang et al.
1991 Sociology Perormance (TIP)
(1991) (2021) companions). A signicant percentage (68 %) o theoretical oundations
Theory o Groups
Gnewuch comes rom the psychology and sociology domain, indicating the sig-
Social Response Nass et al.
1994 Sociology et al. nicance o psychology and sociology in social companionship. Future
Theory (1994)
(2018) researchers may use theories rom communications to develop uturistic
Eagly and Rhee and models or chatbot companions.
1999 Sociology Social Role Theory Wood Choi
(1999) (2020)
Gnewuch 6. Intellectual structure analysis
Computers are Social Nass and
2000 Sociology et al.
Actors (CASA) Moon, 2000
(2018)
The intellectual structure reveals the underlying themes and con-
Aristotle's Theory o Aristotle Bosch et al.
2009 Sociology
Friendship (2009) (2022)
structs building the research domain. Techniques like co-occurrence
Theory o Krämer Payr analysis and bibliographic coupling discern the domain's past, present,
2011 Sociology
Companions et al. (2011) (2011) and uture research directions.
Media and communication
6.1. Keyword co-occurrence analysis (RQ4)
Keyword co-occurrence analysis reveals themes converging the
9
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
research domain (Donthu et al., 2021a; Zupic and Cater, 2015; Callon Table 7
et al., 1983). Fig. 4 presents the co-occurrence network o keywords Thematic Structure o Keyword Co-occurrence Network.
emerged rom VOS viewer analysis. Five thematic clusters emerged, Themes and keywords DG OC APY AC
wherein cluster 1 denotes personication o conversational agents,
Cluster 1 (Red): personication o conversational agents
cluster 2 encapsulates articial companions and Socialbots, cluster 3 Embodied conversational agent 5 8 2014.13 31.25
captures human relations with conversational agents, cluster 4 refects Aective computing 6 6 2011.83 17.33
enablers o conversational agents, and cluster 5 signies AI as social Personication 8 3 2018.33 57.33
companions. Co-occurrence network parameters include occurrence Amazon Echo 7 3 2018.33 55.66
Amazon Alexa 7 3 2019.67 34
(OC), degree (DG), average publication year (APY), and average citation
(AC), as represented in Table 7, to derive an objective assessment o each
Cluster 2 (green): articial companions and social bots
thematic cluster
Conversational agents 23 15 2019.73 9.2
Where OC denotes the requency o a keyword, DG signies the Human-computer interaction 10 6 2020.5 5.66
number o connections that a node (keyword) o the network has with User experience 7 3 2020.33 17
other nodes, APY refects the hotness (more recent) and coldness (less Mobile phone 5 3 2020.67 14
recent) o a keyword, and AC denotes the average impact o that Socialbots 1 3 2019 12.66
Cluster 3 (blue): human relations with conversational agents
keyword on the eld (Donthu et al., 2021a; Lim et al., 2022; Chandra
Articial companions 4 7 2015 13.57
et al., 2022). Robots 6 5 2018.8 9.6
Loneliness 10 5 2020.2 2.6
6.1.1. Cluster 1: personication o conversational agents Older adults 6 4 2020.25 15.5
Friendship 5 3 2020.33 11.66
The rst cluster comprises keywords pointing at personication ca-
Cluster 4 (yellow): enablers o conversational agents
pabilities o conversational agents. The cluster highlights AI agents in Chatbot 25 16 2020.13 26.43
various orms, such as “Amazon Alexa”, “Amazon Echo”, and “embodied Anthropomorphism 25 10 2020.5 37.2
conversational agents”. The highest explored keyword in this cluster is Trust 14 7 2020.71 8.28
“embodied conversational agents” (OC: 8), ollowed by “aective Social presence 11 6 2019.83 63.66
Sel-disclosure 10 6 2020.33 13.5
computing” (OC: 6), and “personication” (OC: 3). The keyword with
Voice assistants 9 5 2020.6 10.2
the highest links and citations in the cluster is “personication” (DG: 8; Cluster 5 (purple): AI as social companions
AC: 57.33), which signies that most AI agents are designed with Articial intelligence 36 24 2019.71 18.79
human-like eatures such as name, gender, voice, etc. Another keyword Socialbots 2 3 2016.33 42.66
Social support 7 3 2020 13
in the cluster, “amazon echo”, received high citations (AC: 55.66) signiy
Companion 3 3 2018.67 4.33
preerence or echo as a companion over other AI agents. Personication Machine learning 6 3 2019 2
is hot and trending topic o the cluster (APY: 2018.33–2019.67).
Abbreviations: DG: degree; OC: occurrence; APY: average publication year; AC:
average citation.
6.1.2. Cluster 2: articial companions and social bots
The second cluster, articial companions and social bots, projects AI
as a companion in the orm o social chatbot, such as “conversational
agents,” and “social bots”. The keyword with the highest connections
Fig. 4. Co-word network o SC with CAs' research.
Notes: Cluster 1 (Red) = Personication o conversa-
tional agents.
Cluster 2 (Green) = Articial Companions and Social
Bots.
Cluster 3 (Blue) = Human relations with conversational
agents.
Cluster 4 (Yellow) = Enablers o conversational agents.
Cluster 5 (Purple) = AI as social companions
10
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
and requency in this cluster is “conversational agents” (DG: 23; OC: 15) coupling technique uses documents or network analysis (Donthu et al.,
ollowed by “human-computer interaction” (DG: 10; OC: 6), “user 2021b; Kessler, 1963). Despite having alternative techniques such as co-
experience” (DG: 7; OC: 3), “mobile phone” (DG: 5; OC: 3), and citation analysis, the current study opts or bibliographic coupling
“Socialbots” (DG: 1; OC: 3). However, in terms o the average citations (which groups “citing” publications) and keyword co-occurrence anal-
“user experience” scored highest (AC: 17) ollowed by “mobile phone” ysis (which groups “current” keywords) to refect current trends existing
(AC: 14) and “social bot” (AC: 12.66). The cluster emphasized the in the eld o SC with CAs (Lim et al., 2022; Donthu et al., 2021a).
enhanced user experience with computers capable o interacting with Noteworthily, the results revealed ve signicant thematic clusters
humans. According to the average publication year index, all the key- discussed in the ollowing subsections. Table 8 lists the top ten articles o
words in this cluster are hot and trending (APY: 2019 – 2020.67). each cluster based on total citations.
6.1.3. Cluster 3: human relations with conversational agents 6.2.1. Cluster 1: articial companions and Socialbots
The third cluster, human relations with conversational agents, in- The rst cluster encapsulates research on articial companions and
cludes ve signicant keywords. The cluster encapsulates the nature o AI-powered social bots. The top ten most cited articles in this cluster are
the relationship between AI and humans. It is made up o keywords such Boshma et al. (2013), Floridi (2008), Elyashar et al. (2013), Orabi et al.
as “articial companions,” “robots,” “loneliness,” “older adults,” and (2020), Biundo et al. (2016), Hepp (2020), Porra et al. (2020), Portela
“riendship.” In terms o requency, “articial companions” has gained and Granell-Canut (2017), Siourti et al. (2018), and Portacolone et al.
the highest (OC: 7), ollowed by “robots” (OC: 5), “loneliness” (OC: 5), (2020) with 116, 45, 37, 34, 19, 15, 15, 13, 12 and 9 citations respec-
“older adults” (OC: 4), and “riendship” (OC: 3). Moreover, “older tively. Boshma et al. (2013), the cluster's most cited article (TC: 116),
adults” revealed highest average citations (AC: 15.5) which indicates ound that social media networks can be easily exploited with the help o
that authors are more inclined towards the relationship between older social bots' inltration campaigns with a success rate o up to 80 %.
adults and articial companions. Also, the keyword “loneliness” has the Elyashar et al. (2013) advocated the use o programmable social bots by
highest connections in the cluster (DG: 10). The cluster emphasizes the organizations on social media to build personal relations with users.
role o AI companions in mitigating the eects o loneliness in older However, data privacy programs should check inltration campaigns.
adults. In terms o average publication year, all the keywords o this Orabi et al. (2020) reviewed the literature on detection methods and
cluster are hot and trending (APY: 2015–2020.33). ound that designing a bot detector is challenging when botmasters keep
evolving with new inltration techniques.
6.1.4. Cluster 4: enablers o conversational agents The second shade o the cluster emphasizes on the emergence o
The ourth cluster, user experience with conversational agents, articial companions. Floridi (2008) identies three important roles o
combines six keywords. The cluster highlights the eatures oered by AI articial companions in the uture, as a partner, inormation-based
companions, such as “anthropomorphism,” “sel-disclosure,” “trust,” server, and a memory steward that could simulate human lie. More-
and “social presence.” The most requently explored keyword in this over, the author observed that the moral aspects o AI companions are
cluster is “chatbot” (OC: 16), ollowed by “anthropomorphism” (OC: still unexplored. Porra et al. (2020) argued that eelings, the substance o
10), “trust” (OC: 7), social “presence” (OC: 6), “sel-disclosure” (OC: 6), humanness, must be reserved only or human interaction. AI compan-
and “voice assistant” (OC: 5). In addition, this cluster holds the highest ions can transorm human behaviour and actions. Biundo et al. (2016)
cited keyword o the entire dataset, which is the “social presence” (AC: invites cross-disciplinary researchers to develop companion applications
63.66). Also, keywords with the highest connections in the cluster are in robotics, health, and elderly care, etc. The cluster also includes studies
“chatbot” (DG: 25) and “anthropomorphism” (DG: 25). Researchers see on AI companions such as Portela and Granell-Canut (2017), Siourti
“social presence” and “anthropomorphic design cues” as mandatory et al. (2018), and Portacolone et al. (2020) that investigated user
eatures to attain SC. A chatbot can become a companion or its users i it experience, user acceptance, and user behaviour with articial com-
behaves like humans and refects social presence. In terms o average panions, respectively.
publication year, all the keywords o this cluster are hot and trending
(APY: 2019.83–2020.71). 6.2.2. Cluster 2: personication o conversational agents
The second cluster ocuses on the personication o CAs. The top ten
6.1.5. Cluster 5: AI as social companions most cited articles in this cluster are Lopatovska and Williams (2018),
The th cluster, AI as social companions, comprises ve signicant Pradhan et al. (2019), Gao et al. (2018), Cho et al. (2019), Ta et al.
keywords. The cluster ocuses on designing articial intelligence and (2020), Skjuve et al. (2021), Crolic et al. (2022), Brause and Blank
machine-learning-based CAs that can play the role o companions, (2020), Kim and Choudhury (2021), and Wilson-Nash et al. (2020) with
providing social support to its users. Keywords enlisted in the cluster are 92, 42, 38, 37, 31, 12, 10, 9 and 9 citations respectively. Lopatovska and
“articial intelligence,” “Socialbots,” “social support,” “companion,” Williams (2018) ound that Alexa's personication behaviour charac-
and “machine learning.” The highest requency in the cluster is o terize as mindless politeness, is the most cited article (TC: 92) in the
“articial intelligence” (OC: 24), ollowed by “Socialbots” (OC: 3), “so- cluster. However, some evidence reveals that people consider Alexa, an
cial support” (OC: 3), “companion” (OC: 3), and “machine learning” associate partner. For example, Gao et al. (2018) observed that users do
(OC: 3). In terms o average citations, the “social bot” has achieved the personiy the echo as an assistant, riend, amily member, wie, and
highest score in the cluster (AC: 42.66). “Articial intelligence” as a girlriend. The ndings also revealed that users who personiy echo tend
keyword has the highest linkages in the entire dataset (DG: 36), as it is to develop more positive emotions with Alexa than those who treat it as
the primary enabler or designing AI agents. Indeed, the research on SC a speaker. Another study by Brause and Blank (2020) identied various
with CAs is a new and emerging eld o study. In terms o average use genres or smart speakers: companionship, sleep aid, peace o mind,
publication year, all the keywords o this cluster are hot and trending sel-control and productivity, increased accessibility, health care and
(APY: 2016.33–2020). support, convenience, and entertainment. Pradhan et al. (2019) ound
that participants fuidly moved between objectiying and personiying
6.2. Bibliographic coupling analysis (RQ4) echo instead o categorizing it straightorwardly. A similar investigation
on older adults by Kim and Choudhury (2021) reveals some benets
This section deals with the bibliographic coupling o themes in SC (such as enjoyment and convenience) and challenges (unctional errors
with CAs. The current study attempts to triangulate the clusters in and limited speech technology) o using smart speakers. Anthropomor-
keyword co-occurrence analysis in the previous section with the phic design cues can lead to better engagement.
bibliographic coupling themes (Goodell et al., 2021). The bibliographic Crolic et al. (2022) ound anthropomorphism negatively aects
11
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Table 8 Table 8 (continued )
Bibliographic coupling themes in SC with CAs. Themes and top articles Authors Citations
Themes and top articles Authors Citations
“In A.I., we trust?” The eects o parasocial
Cluster 1: articial companions and social bots interaction and technician versus luddite
Boshma et al. ideological views on chatbot-based customer Youn and Jin (2021) 23
Design and analysis o a social botnet 116
(2013) relationship management in the emerging
Articial intelligence's new rontier: articial “eeling economy.”
Floridi (2008) 45
companions and the ourth revolution Millennials' attitude towards chatbots: an
De Cicco et al.
Homing Socialbots: intrusion on a specic Elyashar et al. experimental study in a social relationship 21
37 (2020)
organization's employee using Socialbots (2013) perspective
Detection o Bots in social media: a systematic Perceiving a mind in a chatbot: eect o mind
Orabi et al. (2020) 34
review perception and social cues on co-presence, Lee et al. (2020a) 20
Companion-technology: an overview Biundo et al. (2016) 19 closeness, and intention to use
Articial companions, social bots, and work How do AI-driven chatbots impact user
bots: communicative robots as research Hepp (2020) 15 experience? examining gratications, Cheng and Jiang
15
objects o media and communication studies perceived privacy risk, satisaction, loyalty, (2020)
Can computer-based human-likeness endanger and continued use
humanness?” – A philosophical and ethical How chatbots' social presence communication
Porra et al. (2020) 15
perspective on digital assistants expressing enhances consumer engagement: the
Tsai et al. (2021) 8
eelings they can't have mediating role o parasocial interaction and
A new riend in our smartphone? Observing dialogue
Portela and Granell-
interactions with chatbots in the search o 13
Canut (2017)
emotional engagement
Cluster 4: social cues o conversational agents
The CaMeLi ramework—A multimodal virtual
Siourti et al. (2018) 12 A taxonomy o social cues or conversational
companion or older adults Feine et al. (2019) 123
agents
Ethical issues raised by the introduction o
Bickmore and Picard
articial companions to older adults with Portacolone et al. Towards caring machines 64
9 (2004)
cognitive impairment: a call or (2020)
Relational agents improve engagement and Bickmore et al.
interdisciplinary collaborations 50
learning in science museum visitors (2011)
Bickmore et al.
Tinker: a relational agent museum guide 41
Cluster 2: personication o conversational agents (2013)
Personication o the Amazon Alexa: BFF or a Lopatovska and Modalities or building relationships with Bickmore and Mauer
92 25
mindless companion Williams (2018) handheld computer agents (2006)
“Phantom riend” or “just a box with Friendship with a robot: children's perception
inormation”: personication and ontological o similarity between a robot's physical and
Pradhan et al. (2019) 42 Sinoo et al. (2018) 24
categorization o smart speaker-based voice virtual embodiment that supports diabetes
assistants by older adults sel-management
Alexa, my love: analyzing reviews o Amazon Can sotware agents infuence human relations?
Gao et al. (2018) 38 Nakanishi et al.
Echo - Balance theory in agent-mediated 24
(2003)
Once a kind riend is now a thing: communities
understanding how conversational agents at Cho et al. (2019) 37 Eectiveness o an empathic chatbot in
De Gennaro et al.
home are orgotten combating adverse eects o social exclusion 21
(2020)
User experiences o social support rom on mood
companion chatbots in everyday contexts: Ta et al. (2020) 31 The impact o interpersonal closeness cues in
thematic analysis text-based healthcare chatbots on attachment Kowatsch et al.
12
My chatbot companion - a study o human- bond and the desire to continue interacting: (2018)
Skjuve et al. (2021) 12
chatbot relationships An experimental design
Blame the Bot: anthropomorphism and anger in Can we be riends with Mitsuku? A longitudinal
Crolic et al. (2022) 10
customer–chatbot interactions study on the process o relationship Croes and Antheunis
11
Externalized domestication: smart speaker ormation between humans and a social (2021)
Brause and Blank
assistants, networks and domestication 10 chatbot
(2020)
theory
Exploring older adults' perception and use o
Kim and Choudhury Cluster 5: articial intelligence with emotional quotient
smart speaker-based voice assistants: A 9
(2021) From Eliza to XiaoIce: challenges and
longitudinal study Shum et al. (2018) 206
opportunities with social chatbots
Introducing the socialbot: a novel touchpoint Wilson-Nash et al.
9 The design and implementation o XiaoIce, an
along the young adult customer journey (2020) Zhou et al. (2020) 94
empathetic social chatbot
When chatbots meet patients: one-year
Cluster 3 user experience with conversational agents prospective study o conversations between Chaix et al. (2019) 42
Living up to the chatbot hype: the infuence o patients with breast cancer and a chatbot
anthropomorphic design cues and I hear you, I eel you”: encouraging deep sel-
Lee et al. (2020c) 32
communicative agency raming on Araujo (2018) 222 disclosure through a chatbot
conversational agent and company How should my chatbot interact? A survey on
Chaves and Gerosa
perceptions social characteristics in human–chatbot 23
(2021)
AI-based chatbots in customer service and their interaction design
Adam et al. (2021) 64
eects on user compliance Eects o cognitive styles on an MSN virtual
Faster is not always better: understanding the learning companion system as an adjunct to Hsieh (2011) 16
Gnewuch et al.
eect o dynamic response delays in human- 50 classroom instructions
(2018)
chatbot interaction The human side o human-chatbot interaction:
Eects o personalization and social role in a systematic literature review o ten years o Rapp et al. (2021) 14
voice shopping: an experimental study on Rhee and Choi research on text-based chatbots
29
product recommendation by a conversational (2020) Designing a chatbot as a mediator or
voice agent promoting deep sel-disclosure to a real Lee et al. (2020b) 14
The impact o chatbot conversational skills on Schuetzler et al. mental health proessional
25
engagement and perceived humanness (2020) The rise o emotion-aware conversational
Mensio et al. (2018) 8
agents: threats in digital emotions
(continued on next page)
12
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Table 8 (continued ) attachment bonds o the users with healthcare chatbots. Agent output
Themes and top articles Authors Citations modalities also aect the human-computer relationship status. Bickmore
and Mauer (2006) identied our dierent modalities (text with no
Chatbot to improve learning punctuation in
Spanish and enhance open and fexible
Vázquez-Cano et al.
6
image, text with a static image, animated and animated with nonverbal
(2021) speech) and ound that embodied and animated agents ormed stronger
learning environments
social bonds with the users. This cluster also highlights the role o AI in
combating loneliness and social exclusion. Interestingly, empathetic
chatbots can now mitigate the eects o social exclusion on mood and
customer satisaction because o loty expectations rom a humanlike
eelings (De Gennaro et al., 2020). Bickmore and Picard (2004)
chatbot and suggested empathetic chatbots can tackle human emotions
demonstrate that computers can signicantly impact users' perception o
(Crolic et al., 2022). Smart speaker usage decreases over time and loses
care. Bickmore et al. (2011) and Bickmore et al. (2013) ound increased
its presence at home (Cho et al., 2019). Notably, most studies examined
engagement o museum visitors when interacting with a virtual
user behaviour with Alexa or Google Home, however, advanced CAs
anthropomorphic robot. However, a study on diabetic children by Sinoo
provide more emotional and social support to their users. For example,
et al. (2018) revealed that physical robots established a stronger
Replika interacts with its users to ull their emotional desires. Ta et al.
riendship with children than virtual avatars. Contrarily, Croes and
(2020) examined 1854 user reviews and interviewed sixty-six o Replika
Antheunis (2021), a study on user relations with “Mitsuku”, ound that
users and ound that Replika can mitigate the eelings o loneliness in
participants experienced low eelings o riendship with the bot; also, the
users. Skjuve et al. (2021) echoed Ta et al. (2020) work and identied
social process decreased ater each interaction.
key characteristics o Replika that are the non-judgmental, under-
standing and accepting nature o the chatbot.
6.2.5. Cluster 5: articial intelligence with emotional quotient
The th cluster concentrates on emotion-aware chatbots that can
6.2.3. Cluster 3: user experience with conversational agents
understand human eelings. The top ten cited articles in this cluster are
The third cluster captures research on user experience with CAs. The
Shum et al. (2018), Zhou et al. (2020), Chaix et al. (2019), Lee et al.
top ten most cited articles in this cluster are Araujo (2018), Adam et al.
(2020c), Chaves and Gerosa (2021), Hsieh (2011), Rapp et al. (2021),
(2021), Gnewuch et al. (2018), Rhee and Choi (2020), Schuetzler et al.
Lee et al. (2020b), Mensio et al. (2018), and Vázquez-Cano et al. (2021)
(2020), Youn and Jin (2021), De Cicco et al. (2020), Lee et al. (2020a),
with 206, 94, 42, 32, 23, 16, 14, 14, 8, 6 respectively. Shum et al. (2018)
Cheng and Jiang (2020), Tsai et al. (2021) with 222, 64, 50, 29, 25, 23,
review o chatbot evolution rom Eliza (1960) to Xiaoice (2014), which
21,20,15, and 8 citations respectively. The most cited article (TC: 222)
demonstrated how XiaoIce could engage humans or long conversations
in the cluster, Araujo (2018), revealed that anthropomorphic design
through recognizing their emotions, has received the highest citations
cues infuence consumers' emotional connection with the organization
(TC: 206) in the cluster. Another study on XiaoIce by Zhou et al. (2020)
with a mediating role o social presence. AI-driven chatbots oer utili-
measured the chatbot's eectiveness using conversation turns per ses-
tarian, hedonic, technology, and social gratications that lead to
sion (CPS) and ound that the chatbot achieved an average CPS o 23,
customer satisaction, customer loyalty, and continued use (Cheng and
higher than any other chatbot or even humans. Chatbots' sel-disclosure
Jiang, 2020). The purpose o anthropomorphism is to oer a human-like
can trigger humans to disclose their personal eelings and thoughts (Lee
interaction experience to consumers. Gnewuch et al. (2018) ound that
et al., 2020c). “Conucius”, a virtual learning companion, signicantly
chatbots' dynamic delayed responses increased perceived humanness
beneted eld-dependent learners (Hsieh, 2011). Even in the eld o
and customer satisaction. A similar study by Adam et al. (2021) ound
language learning, students value chatbots as it provides greater support
that both anthropomorphism and consistent staying o the chatbot can
and companionship in the learning process (Vázquez-Cano et al., 2021).
signicantly increase user compliance on chatbot requests. Schuetzler
All humans need an understanding and supporting associate every-
et al. (2020) demonstrated that the conversational skill o the chatbot
where, irrespective o the eld (teaching, health and support, enter-
leads to anthropomorphism and social presence.
tainment, etc.). Empathetic conversational agents can now do the
Moreover, instead o an assistant, a riend role played by the chatbot
needul with the help o an inbuilt emotional quotient. Chaix et al.
positively infuences consumer engagement. De Cicco et al. (2020)
(2019) conducted a study on 4797 cancer patients and observed that 88
showed that instead o being task-oriented, social-oriented interaction
% o participants elt that “Vik”, a social chatbot helped and supported
style o chatbot positively infuences user's perception o social presence.
them in tracking their treatment. A similar study by Lee et al. (2020b)
Also, the social role played by chatbots and personalized interaction
ound that participants revealed more inormation to a chatbot than a
infuence customer attitudes towards the product recommended by the
mental health proessional. Considering the development o conversa-
conversational agent (Rhee and Choi, 2020). When interacting with
tional agents to an emotional awareness level, Mensio et al. (2018)
media characters, the user's interpersonal involvement represents par-
questioned the understanding o chatbots towards human values.
asocial interaction. Perceived parasocial interaction mediates the in-
fuence o social presence on consumer engagement outcomes (Tsai
7. Towards a conceptual framework (RQ5)
et al., 2021). Additionally, Youn and Jin (2021) ound that a riend
chatbot can build stronger parasocial interactions with consumers
The current study presents the triangulation o major themes in SC
compared to an assistant chatbot.
with CAs using keyword co-occurrence analysis and bibliographic
coupling analysis (Table 9). Extant literature is scant in presenting a
6.2.4. Cluster 4: social cues o conversational agents
holistic view on social companionship with AI. Thus, this study proposes
The ourth cluster highlights research on the social cues o conver-
a conceptual ramework that has emerged rom the content analysis o
sational agents. The top ten most cited articles in this cluster are Feine
the articles used or conducting the current review. The ramework
et al. (2019), Bickmore and Picard (2004), Bickmore et al. (2011),
details the antecedents, mediators, moderators, and outcomes o estab-
Bickmore et al. (2013), Bickmore and Mauer (2006), Sinoo et al. (2018),
lishing SC with CAs presented in Fig. 5, which provides oundational
Nakanishi et al. (2003), De Gennaro et al. (2020), Kowatsch et al.
knowledge to uture researchers and scholars in SC with CAs. Note-
(2018), and Croes and Antheunis (2021) with 123, 64, 50, 41, 25, 24,
worthily, although the ramework presents a comprehensive view o SC
24, 21, 12, and 11 respectively. Feine et al. (2019) review segmented
with CAs, this should not be considered as specic or all-inclusive;
orty-eight social cues into our categories (verbal, visual, auditory, and
rather, it should be viewed as a source o elemental knowledge that
invisible). Kowatsch et al. (2018) identied closeness cues (such as vi-
practitioners, designers, managers, business owners, and uture re-
sual, verbal, quasi-nonverbal, and relational) that can infuence the
searchers can utilize to expand the existing boundaries o the domain.
13
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Table 9 Table 9 (continued )
Summary o research on SC with CAs. Perormance Keyword co- Bibliographic Conceptual
Perormance Keyword co- Bibliographic Conceptual analysis occurrence analysis coupling ramework
analysis occurrence analysis coupling ramework
loneliness reduction,
• Field • Body o emotional
• Body o • Body o
perormance knowledge connection,
knowledge knowledge
• Bibliometric • Citing continued use,
• Author keywords • Based on actors
inormation publication service loyalty
Perormance Themes Themes Factors
7.1. Antecedents
Antecedents o SC
with CA
Antecedents are determinants o the outcomes, and their eects on
Conversational the outcome variables can be mediated and moderated by intervening
capability: variables. The current research broadly buckets antecedents o SC with
conversational skill,
humour and voice,
CAs into three categories, conversational capability and unctional
interaction style, capability. Conversational capability refects the aspects related to skills
response time, sel- o articial intelligence to communicate and interact with humans, such
expressive as conversational skill (Schuetzler et al., 2020), humour and voice
Functional
(Moussawi and Benbunan-Fich, 2021), interaction Style (De Cicco et al.,
Publication Cluster1: capability:
activity: a personication o inormation and 2020), the response time (Gnewuch et al., 2018) and sel-expressivity
total o 126 conversational entertainment (Ramadan, 2021). Functional capability reers to the aspects related to
articles were agents: embodied Social capability: the skills o AI to perorm various tasks or the users, such as providing
published conversational role o chatbot, inormation (Cheng and Jiang, 2020) and entertainment (Cheng and
between agent, aective anthropomorphic
2003 and computing, design, media
Jiang, 2020). Social Capability refects the aspects related to skills o AI to
May 2022. personication, richness and mind provide social support and connection to the users such as the role o
Amazon Echo, perception chatbot (Youn and Jin, 2021), anthropomorphic design cues (Araujo,
Top sources Amazon Alexa Cluster 1: 2018), media richness (Hsieh and Lee, 2021), mind perception (Lee
articial Mediators or SC
et al., 2020a).
• Most Cluster 2: articial companions with CAs:
citations: companions and and Socialbots perceived dialogue,
computers in Social Bots: trust attitude, 7.2. Mediators
human conversational Cluster 2: product
behaviour agents, human- personication involvement,
(TC: 286 computer o perceived Mediators are intervening actors that dene the nature o relation-
citations) interaction, user conversational intelligence, ship between antecedents and outcomes. The study identied mediators
• Most experience, mobile agents perceived such as perceived dialogue (Tsai et al., 2021), trust (Pitardi and Mar-
publications: phone, Socialbot anthropomorphism, riott, 2021), attitude (Hsieh and Lee, 2021), product involvement (Rhee
lecture notes Cluster 4: social social presence,
in computer Cluster 3: human cues o parasocial
and Choi, 2020), perceived intelligence (Moussawi and Benbunan-Fich,
science (NP: relations with conversational interaction, privacy 2021), perceived anthropomorphism (Moussawi and Benbunan-Fich,
7 conversational agents concern, task and 2021), social presence (Adam et al., 2021), parasocial Interaction
publications) agents: articial social attraction, (Youn and Jin, 2021), privacy Concern (Bawack et al., 2021), task and
companions, robots, Cluster 3: user closeness, love,
social attraction (Lei et al., 2021), closeness (Lee et al., 2020a), love
Top loneliness, older experience passionate desire to
publications: adults, riendship with use AI (Hernandez-Ortega and Ferreira, 2021), and passionate desire to use AI
Araujo, 2018 conversational (Ramadan, 2021).
(222 citations) Cluster 4: enablers agents Moderators or SC
o conversational with CA
Top authors agents: Chatbot, Cluster 5: situational 7.3. Moderators
anthropomorphism, articial characteristics:
• Most trust, social intelligence (brand involvement, Moderators infuence the degree o relationship between antecedents
citations: presence, sel- with emotional mood - angry or and outcomes; antecedents and mediators; and mediators and conse-
Shum HY and disclosure, voice quotient. calm)
quences. Two moderators play the oremost role in dening social
Li D (three assistants
hundred User characteristics: companionship with conversational agents: situational actors and user
citations Cluster 5: AI as psychographic characteristics. Situational characteristics refect the environmental in-
each) social companions: characteristics fuence, such as brand involvement (Hasan et al., 2021) and customer
• Most articial (experience with
mood (angry or Calm) (Crolic et al., 2022), while user characteristics
publications: intelligence, chatbot, consumer
Bickmore T Socialbots, social innovativeness, refect the individual's psychographic actors, such as past experience
(our support, companion, ideological views (Kowatsch et al., 2018), consumer innovativeness (Kasilingam, 2020),
publications) machine learning and ideological views on technology (Youn and Jin, 2021), as well as
Demographic demographic characteristics o individuals such as their age (Pradhan
characteristics: age,
et al., 2019) and gender (Bergen, 2016).
gender
Outcomes o SC with 7.4. Outcomes
CA:
intention to use,
user engagement, Outcomes are the actors that results rom the infuence o ante-
perceived cedents and intervening variables like mediators and moderators. Social
humanness, user companionship with conversational agents lead to intention to use
experience,
chatbots (Lee et al., 2020a), user engagement (Tsai et al., 2021),
perceived humanness (Gnewuch et al., 2018), user experience (Bawack
14
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Fig. 5. Conceptual ramework or antecedents, mediators, moderators, and outcomes o SC with CAs.
et al., 2021), loneliness reduction (Jones et al., 2021), emotional are now developing AI agents to orm emotional connections with
connection (Araujo, 2018), continued use (Cheng and Jiang, 2020), and humans. The multiold applications make the conversational commerce
service Loyalty (Hernandez-Ortega and Ferreira, 2021). vital and uturistic warranting academicians and practitioners'
The proposed conceptual ramework unurls three oundational ca- attention.
pabilities in a conversational AI that intends to orm a relational bond In terms o the source perormance, review ound that Computers in
with the users: Conversational Capabilities, Functional Capabilities and Human Behaviour is the preerred avenue or publication with the
Social Capabilities. Future researchers are guided to investigate each highest citations (286). Majorly the top sources published in the eld
capability and its signicance in enabling CA designs to achieve elevated belongs to computer science, communication, human-computer inter-
companionship levels. As the ramework emerges rom the extant action, and psychology, highlighting the multidisciplinary scope or
literature, one notable observation is that there can be an addition to research on SC with CAs. Regarding the author's perormance, the study
capabilities beyond what is already in the literature, such as emotional, ound that Shum and Li are the most cited (300 citations) authors in the
motivational, and educational capabilities. This opens immense oppor- eld. However, Bickmore has the highest number o articles (4). The top
tunities or practitioners to design AI companions with multiple capa- authors come rom both practice and academics. Moreover, signicant
bilities. The ramework also proposes that the more capable a CA is, the contributions come rom western countries, which suggests more
stronger the companionship it can develop with the user. investigation and research are required in other parts o the globe to
promote higher diversity and inclusivity in the eld. The study has also
8. Disscussion revealed a plethora o theories used in SC with CAs. Theories are rom
varied disciplines such as sociology, psychology, media and communi-
The present study oered a state-o-the-art literature review cation, marketing, and computer science, with the earliest theory rom
encompassing science mapping and intellectual structure analysis. Sci- the eld o psychology in 1946 (i.e., balance theory) and the latest
ence mapping entail perormance analysis o scientic actors like theory rom the eld o sociology in 2018 (i.e., actor-network theory).
sources, authors and documents in the research domain. Intellectual Noteworthily, psychology and sociology are prominent as the area is
structure analysis entail keyword co-occurrence and bibliographic primarily associated with human psychology and relationship develop-
coupling analysis. Intellectual structure analysis also discerns major ment. Future researchers are encouraged to use communication the-
theories, themes and conceptual ramework encompassing antecedents, ories, morality and ethics, and theories rom human relations, which can
mediator, moderators, and outcomes o social companionship with help develop and examine more advanced AI companions.
conversational commerce.
The science mapping reveals the growing interest in the research 8.1. Theoretical contributions and implications
domain, and the spurt indicates the relevance o the topic in the
contemporary times. The study identied 126 articles on SC with CAs The present study's theoretical contributions align with the guide-
published between 2003 and May 2022 (RQ1). The publication trend o lines or advancing theories with the help o bibliometric research
the eld indicated that about 66 % o articles have appeared only in the (Mukherjee et al., 2022). The study draws three essential and notable
last three years (2019–2022). The recent emergence o the topic is due to contributions. The rst contribution appears to promote the objective
the technological advancement through AI in the ourth industrial rev- discovery o the eld's knowledge clusters, namely articial companions
olution that gave rise to several AI applications in the market. Businesses and Socialbots, personication o conversational agents, user experience
15
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
with conversational agents, social cues o conversational agents, and adopting advanced technologies (Omrani et al., 2022). Future re-
articial intelligence with emotional quotient. The knowledge clusters searchers and practitioners must design companions laden with moral
enrich the understanding o research design and philosophy along with values and ethics to trigger positive thoughts in human minds. Thus, the
the major streams. The objective discovery also helps strengthen and study proposes the ollowing research questions:
develop the domain's theoretical oundations. More importantly, uture
FRQ1. How does social bot inltration infuence consumer behaviour?
researchers can identiy and explore the research stream to ll the
existing research gaps discussed in a later section o this paper. The FRQ2. What are articial companions' positive and negative impacts on
second contribution appears to deliver a nomological clarity o con- human emotions, decisions, and actions?
structs and the network that maps them together through co-occurrence
FRQ3. What actors are responsible or perceiving an articial companion
analysis o author keywords. Finally, the third contribution comes
as moral and ethical?
through the development o the conceptual ramework (Fig. 5), along
with enlisting the theories in SC with CAs (Table 6). The proposed ho-
8.3.2. Personication o conversational agents
listic conceptual ramework refects the eld's antecedents, mediators,
Anthropomorphic chatbot design leads to personication behaviour
moderators, and outcomes, unolding numerous opportunities or
but, at times, increases user expectations, which may negatively aect
theoretical developments to the current understanding o the domain.
customer satisaction when a customer is angry (Crolic et al., 2022).
Future researchers should design empathetic chatbots to understand
8.2. Managerial contributions and implications
user emotions beore responding to users' needs. Future bots should
redirect clients to human executives i they sense uriated customers.
The managerial contribution o the current study aligns practice with
Also, AI can help patients and older adults remember the medicines and
theory (Mukherjee et al., 2022). First, this review contributes to practice
diet plans to ensure their well-being. Extant literature points at chatbot
by providing managers with a macroscopic overview o research and
personality; however, it is silent on the infuence o the user's personality
development in social companionship with conversational agents. The
on human-computer interaction. Extrovert users nd AI agents as talk-
top publications in Table 4 can help managers know more about arti-
ative and active compared to introverts. Users may lose connection with
cial companions and their relationships with humans. More specically,
the smart speakers due to the conversational agent's inability to modiy
the review allows managers to look at dierent streams o research on
their conversation style based on the user's personality (Cho et al.,
social companionship with AI, namely articial companions and
2019). Future researchers should work on CAs that assess users' per-
Socialbots, personication o conversational agents, user experience
sonalities and interact accordingly. Thus, the study proposes the
with conversational agents, social cues o conversational agents, and
ollowing uture research questions:
articial intelligence with emotional quotient along with top publica-
tions o each stream. FRQ4. What actors are responsible or perceiving an articial agent with a
Second, this study contributes to practice by presenting the list o top caring and loving personality?
authors in the eld o SC with CAs, which can help managers to reach
FRQ5. How should conversational agents be designed that support patients'
out to the experts in the area or any guidance they require while
well-being, lonely individuals, and older citizens?
designing or developing a new articial companion. Managers and
practitioners can contact the top authors to gain an expert opinion on SC FRQ6. What is the impact o user personality on interaction with articial
with CAs using Table 5. Though experts and leading authors are oten companions?
busy and are not readily available or casual appointments, they
consider good collaboration opportunities oered respectully depend- 8.3.3. User experience with conversational agents
ing upon the research model, scope, and interest. User experience has always taken center stage whenever technology
Third, this study contributes to practice by proposing the conceptual transorms the business. According to Araujo (2018), the anthropo-
ramework (Fig. 5) developed through the content analysis o 126 arti- morphic design o chatbots lead to an emotional connection between the
cles that can help managers to look at antecedents, mediators, moder- organization and users. The stream highlights research on user percep-
ators, and outcomes o SC with CAs. Also, managers are encouraged to tion, satisaction, and experience with conversational agents based on
consider essential variables while designing companion technology or the perception o social presence and anthropomorphism. However, the
the marketplace. literature is scant on the role o SC in infuencing the user experience.
Factors such as perceived intelligence, anthropomorphism, and social
8.3. Future research directions (RQ6) presence are predominant in the literature, while conversational capa-
bility's role in building relationships remained unexplored. Schuetzler
The review suggests several uture research agendas or the major et al. (2020) ound that conversational skill leads to perceived anthro-
themes identied in the study on SC with CAs. The ollowing sub- pomorphism and social presence. Future researchers should study the
sections unold new and exciting avenues in the domain. determinants o conversational capability and its infuence on human-
computer relationships. Communication theories could help uture re-
8.3.1. Articial companions and Socialbots searchers design more interesting articial companions. Thereore, the
Empathetic CAs oer companionship to humans. However, social study proposes the ollowing research questions:
bots may exploit and misuse users' personal inormation via inltration
FRQ7. What is the user experience in a relationship with articial
on social media platorms such as Facebook and Twitter (Boshma et al.,
companions?
2013; Elyashar et al., 2013). Social bot design advancements challenge
bot detection (Orabi et al., 2020). It leads to conusion in the emerging FRQ8. Does human-computer interaction aect the user's relations with
metaverse to gure out whether proles are o real humans or ake bots. other humans in society?
Future researchers need to design ecient socialbot detectors to deal
FRQ9. What actors determine the conversational capability o an articial
with this conusion. AI companions also bring challenges; or example,
companion?
the growing infuence o such technology in human society will reshape
human emotions, decisions, and actions (Floridi, 2008). AI companions
8.3.4. Social cues o conversational agents
may ampliy the negative emotions o the user i not programmed
Conversational agents display various social cues, such as verbal,
morally. Negative emotions may induce depression and suicidal ten-
visual, auditory, and invisible (Feine et al., 2019). These social cues
dencies (Possati, 2022). Trust, is always a concern or users beore
16
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
create warmth in the conversational agents, making them converse like CRediT authorship contribution statement
humans. Bickmore and Mauer (2006) ound that embodied and
animated relational agents tend to orm stronger social bonds with the All authors have contributed to all aspects o this paper.
users than disembodied and non-animated agents. Thus, it would be
interesting to know about the social signals responsible or establishing
an emotional connection between users and CAs. Croes and Antheunis Declaration of competing interest
(2021) ound that social processes decrease ater interacting with the
chatbot. However, studies on chatbots like Tinker and Replika suggest There is no conficts o interest or this manuscript.
that users can orm long-time relationships with AI companions (Bick-
more et al., 2011; Bickmore et al., 2013; Ta et al., 2020; Skjuve et al., Data availability
2021). The contradictory ndings suggest uture researchers explore the
comparative perormance o two or more articial companions, such as Data will be made available on request.
Mitsuku and Replika. Thus, the study proposes the ollowing uture
research questions: References
FRQ10. Which social signals or cues are responsible or establishing Aaker, J.L., 1997. Dimensions o brand personality. J. Mark. Res. 34 (3), 347–356.
companionship and emotional connection with the conversational agents? Adam, C., Cavedon, L., Padgham, L., 2010, July. “Hello Emily, how are you today?”-
personalized dialogue in a toy to engage children. In: Proceedings o the 2010
FRQ11. What is the comparative perormance o two or more articial Workshop on Companionable Dialogue Systems, pp. 19–24.
companions? Adam, M., Wessel, M., Benlian, A., 2021. AI-based chatbots in customer service and their
eects on user compliance. Electron. Mark. 31 (2), 427–445.
FRQ12. How do service settings, socioeconomic status and cultural norms Aksnes, D.W., Langeldt, L., Wouters, P., 2019. Citations, citation indicators, and
research quality: an overview o basic concepts and theories. SAGE Open 9 (1),
infuence the selection o social cues and design o articial companions? 2158244019829575.
Altman, I., Taylor, D.A., 1973. Social Penetration: The Development o Interpersonal
8.3.5. Articial intelligence with emotional quotient Relationships. Holt, Rinehart & Winston.
Ameen, N., Tarhini, A., Reppel, A., Anand, A., 2021. Customer experiences in the age o
Conversational agents have evolved through the years, rom ELIZA
articial intelligence. Comput. Hum. Behav. 114, 106548.
(1960) to empathetic chatbots like XiaoIce (2014) (Shum et al., 2018). Araujo, T., 2018. Living up to the chatbot hype: the infuence o anthropomorphic design
Empathetic chatbots are conversational agents that can recognize and cues and communicative agency raming on conversational agent and company
understand human emotions. Recently, Google's Lambda is believed to perceptions. Comput. Hum. Behav. 85, 183–189.
Aristotle (2009) Nicomachean Ethics (trans: Ross, W.D.). Oxord: Oxord University.
have become sentient and developed human-like eelings (that include Aw, E.C.X., Tan, G.W.H., Cham, T.H., Raman, R., Ooi, K.B., 2022. Alexa, what’s on my
love, ear, condence, etc.). Noteworthily the degree o chatbot's sel- shopping list? Transorming customer experience with digital voice assistants.
awareness or emotional awareness is unexplored. It opens new ave- Technol. Forecast. Soc. Chang. 180, 121711.
Barrett, K., Campos, J., 1987. Perspectives on emotional development: II. A unctionalist
nues o research on articial lie, articial consciousness, and articial approach to emotions. In: Handbook o Inant Development, 2nd ed. Wiley-
beings. There is scope or uture research comparing human empathy Interscience, New York, pp. 555–578.
with AI empathy. Articial intelligence with an intellectual quotient has Bawack, R.E., Wamba, S.F., Carillo, K.D.A., 2021. Exploring the role o personality, trust,
and privacy in customer experience perormance during voice shopping: evidence
snatched several jobs rom humans in the market. Would it be accept- rom SEM and uzzy set qualitative comparative analysis. Int. J. In. Manag. 58,
able to design AI with an emotional quotient too? And what about hu- 102309.
manity, then? Feelings and emotions are humans' assets, making them Bem, D.J., 1972. Sel-perception theory. Adv. Exp. Soc. Psychol. 6, 1–62.
Benyon, D., Mival, O., 2010, December. From human-computer interactions to human-
dierent rom machines. As Porra et al. (2020) questioned the companion relationships. In: Proceedings o the First International Conerence on
Humanlikliness o AI agents, arguing that eelings, the very substance o Intelligent Interactive Technologies and Multimedia, pp. 1–9.
humanness, must be reserved only or human interaction. Thereore, the Bergen, H., 2016. ‘I'd blush i I could': digital assistants, disembodied cyborgs and the
problem o gender. Word and text, A Journal o Literary Studies and Linguistics, 6
study proposes the ollowing research questions:
(01), 95–113. In: Proceedings o the 8th International Conerence on Human-Agent
Interaction, pp. 221–223.
FRQ13. What actors indicate that a conversational agent has eelings,
Bickmore, T., Mauer, D., 2006, April. Modalities or building relationships with handheld
sel-awareness, and consciousness like humans? computer agents. In: CHI'06 Extended Abstracts on Human Factors in Computing
Systems, pp. 544–549.
FRQ14. How do individual dierences aect their perceived awareness Bickmore, T., Peier, L., Schulman, D., 2011, September. Relational agents improve
and consciousness o a conversational agent? engagement and learning in science museum visitors. In: International Workshop on
Intelligent Virtual Agents. Springer, Berlin, Heidelberg, pp. 55–67.
FRQ15. What are the opportunities, challenges, and threats to human so- Bickmore, T.W., Picard, R.W., 2004, April. Towards caring machines. In: CHI’04
ciety rom articial intelligence with an emotional quotient? Extended Abstracts on Human Factors in Computing Systems, pp. 1489–1492.
Bickmore, T.W., Vardoulakis, L.M.P., Schulman, D., 2013. Tinker: a relational agent
museum guide. Auton. Agent. Multi-Agent Syst. 27 (2), 254–276.
9. Conclusion Biundo, S., Höller, D., Schattenberg, B., Bercher, P., 2016. Companion-technology: an
overview. Künstl. Intell. 30 (1), 11–20.
Bloch, L.-R., Lemish, D., 1999. Disposable love: the rise and all o a virtual pet. New
This study is a comprehensive systematic review o SC with CAs that Media Soc. 1 (3), 283–303.
attempts to reveal the nuances o articial companions and their rela- Blondel, V.D., Guillaume, J.L., Lambiotte, R., Leebvre, E., 2008. Fast unolding o
tionship with humans. The covid-19 pandemic and other reasons or communities in large networks. J. Stat. Mech: Theory Exp. 10008, 1–12.
Bosch, M., Fernandez-Borsot, G., Comas, M.I., Figa Vaello, J., 2022. Evolving riendship?
social exclusion result in eelings o loneliness among individuals, Essential changes, rom social networks to articial companions. Soc. Netw. Anal.
including older adults. The articial intelligence expertise o human Min. 12 (1), 1–10.
society has overcome this issue by designing advanced articial com- Boshma, Y., Muslukhov, I., Beznosov, K., Ripeanu, M., 2013. Design and analysis o a
social botnet. Comput. Netw. 57 (2), 556–578.
panions such as XiaoIce and Replika, which can unction as therapeutic Bothun, D., Lieberman, M., Rao, A.S., 2017. Bot. In: Me: A Revolutionary Partnership.
resources or their users. The uture o human-computer interaction lies Bourdieu, P., 1980. Le sens pratique. Minuit, Paris.
in developing AI companions, their capabilities, determinants, social Boyack, K.W., Klavans, R., 2010. Co-citation analysis, bibliographic coupling, and direct
citation: which citation approach represents the research ront most accurately.
acceptability, and their infuence on society. The study highlights the
J. Am. Soc. In. Sci. Technol. 61 (12), 2389–2404.
usage o articial companions in each service sector (hospitality, Bracken, L.J., Oughton, E.A., 2006. ‘What do you mean?’The importance o language in
tourism, education, healthcare, entertainment, etc.). Thereore, AI developing interdisciplinary research. Trans. Inst. Br. Geogr. 31 (3), 371–382.
companions in the customer journey are a paradigm shit in designing Brause, S.R., Blank, G., 2020. Externalized domestication: smart speaker assistants,
networks and domestication theory. In. Commun. Soc. 23 (5), 751–763.
the marketing strategies. In such a scenario, this literature review oers Burgoon, J.K., Hale, J.L., 1988. Nonverbal expectancy violations: model elaboration and
a bird's eye view on the emerging eld o SC with CAs. application to immediacy behaviors. Commun. Monogr. 55 (1), 58–79.
17
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Callon, M., Courtial, J.P., Turner, W.A., Bauin, S., 1983. From translations to problematic Gao, Y., Pan, Z., Wang, H., Chen, G., 2018, October. Alexa, my love: analyzing reviews o
networks: an introduction to co-word analysis. Soc. Sci. In. 22 (2), 191–235. amazon echo. In: 2018 IEEE Smart World, Ubiquitous Intelligence and Computing,
Campos, J., Paiva, A., 2010, September. May: my memories are yours. In: International Advanced and Trusted Computing, Scalable Computing and Communications, Cloud
Conerence on Intelligent Virtual Agents. Springer, Berlin, Heidelberg, pp. 406–412. and Big Data Computing, Internet o People and Smart City Innovation. IEEE,
Cavazza, M., Smith, C., Charlton, D., Zhang, L., Turunen, M., Hakulinen, J., 2008. pp. 372–380.
A ‘companion’ ECA with planning and activity modelling. Auton. Agent. Multi-Agent Gasteiger, N., Loveys, K., Law, M., Broadbent, E., 2021. Friends rom the uture: a
Syst. 1281–1284. scoping review o research into robots and computer agents to combat loneliness in
Chaix, B., Bibault, J.E., Pienkowski, A., Delamon, G., Guillemassé, A., Nectoux, P., older people. Clin. Interv. Aging 941–971.
Brouard, B., 2019. When chatbots meet patients: one-year prospective study o Gnewuch, U., Morana, S., Adam, M. T., and Maedche, A. (2018). Faster is not always
conversations between patients with breast cancer and a chatbot. JMIR Cancer 5 (1), better: understanding the eect o dynamic response delays in human-chatbot
e12856. interaction. In 26th European Conerence on Inormation Systems: Beyond
Chandra, S., Verma, S., Lim, W.M., Kumar, S., Donthu, N., 2022. Personalization in Digitization-acets o Socio-technical Change, ECIS 2018, Portsmouth, UK, June 23-
personalized marketing: trends and ways orward. Psychol. Mark. 39 (8), 28, 2018. Ed.: U. Frank (p. 143975).
1529–1562. Goodell, J.W., Kumar, S., Lim, W.M., Pattnaik, D., 2021. Articial intelligence and
Chaturvedi, R., Verma, S., 2022. Articial Intelligence-Driven Customer Experience: machine learning in nance: identiying oundations, themes, and research clusters
Overcoming the Challenges. Caliornia Management Review Insights. Accessed rom. rom the bibliometric analysis. J. Behav. Exp. Financ. 32, 100577.
https://cmr.berkeley.edu/assets/documents/pd/2022-03-artiicial-intelligence-dr Ha, Q.A., Chen, J.V., Uy, H.U., Capistrano, E.P., 2021. Exploring the privacy concerns in
iven-customer-experience-overcoming-the-challenges.pd. using intelligent virtual assistants under perspectives o inormation sensitivity and
Chaves, A.P., Gerosa, M.A., 2021. How should my chatbot interact? A survey on social anthropomorphism. Int. J. Hum. Comput. Interact. 37 (6), 512–527.
characteristics in human–chatbot interaction design. Int. J. Hum. Comput. Interact. Hamilton, R., Ferraro, R., Haws, K.L., Mukhopadhyay, A., 2021. Traveling with
37 (8), 729–758. companions: the social customer journey. J. Mark. 85 (1), 68–92.
Chen, C., Ibekwe-SanJuan, F., Hou, J., 2010. The structure and dynamics o co-citation Hasan, R., Shams, R., Rahman, M., 2021. Consumer trust and perceived risk or voice-
clusters: a multiple perspective co-citation analysis. J. Am. Soc. In. Sci. Technol. 61 controlled articial intelligence: the case o Siri. J. Bus. Res. 131, 591–597.
(7), 1386–1409. Heider, F., 1946. Attitudes and cognitive organization. J. Psychol. 21 (1), 107–112.
Cheng, Y., Jiang, H., 2020. How do AI-driven chatbots impact user experience? Hepp, A., 2020. Articial companions, social bots and work bots: communicative robots
Examining gratications, perceived privacy risk, satisaction, loyalty, and continued as research objects o media and communication studies. Media Cult. Soc. 42 (7–8),
use. J. Broadcast. Electron. Media 64 (4), 592–614. 1410–1426.
Cho, M., Lee, S.S., Lee, K.P., 2019, June. Once a kind riend is now a thing: Hernandez-Ortega, B., Ferreira, I., 2021. How smart experiences build service loyalty:
Understanding how conversational agents at home are orgotten. In: Proceedings o the importance o consumer love or smart voice assistants. Psychol. Mark. 38 (7),
the 2019 on Designing Interactive Systems Conerence, pp. 1557–1569. 1122–1139.
Cialdini, R.B., 2001. Harnessing the Science o Persuasion. Hirano, T., December 2016. Gatebox, Holographic Virtual Assistant, Launches Pre-orders
Cole, T., Bradac, J.J., 1996. A lay theory o relational satisaction with best riends. or Geeks in Japan, US. The Bridge. Retrieved rom. http://thebridge.jp/en/2016/1
J. Soc. Pers. Relat. 13 (1), 57–83. 2/gatebox-launch. Retrieved rom.
Conversational AI market report (2021). Markets and markets.com. Accessed on Hjørland, B., 2013. Citation analysis: a social and dynamic approach to knowledge
February, 27 2023, at https://www.marketsandmarkets.com/Market-Reports/c organisation. In. Process. Manag. 49 (6), 1313–1325.
onversational-ai-market-49043506.html#:~:text=%5B317%20Pages%20Report% Homans, G.C., 1958. Social behavior as exchange. Am. J. Sociol. 63 (6), 597–606.
5D%20The%20global,USD%206.8%20billion%20in%202021. Hsieh, S.H., Lee, C.T., 2021. Hey Alexa: examining the eect o perceived socialness in
Croes, E.A., Antheunis, M.L., 2021. Can we be riends with Mitsuku? A longitudinal study usage intentions o AI assistant-enabled smart speaker. J. Res. Interact. Mark. 15 (2),
on the process o relationship ormation between humans and a social chatbot. 267–294.
J. Soc. Pers. Relat. 38 (1), 279–300. Hsieh, S.W., 2011. Eects o cognitive styles on an MSN virtual learning companion
Crolic, C., Thomaz, F., Hadi, R., Stephen, A.T., 2022. Blame the bot: anthropomorphism system as an adjunct to classroom instructions. J. Educ. Technol. Soc. 14 (2),
and anger in customer–chatbot interactions. J. Mark. 86 (1), 132–148. 161–174.
Dat, R.L., Lengel, R.H., 1986. Organizational inormation requirements, media richness Huang, M.H., Rust, R., Maksimovic, V., 2019. The eeling economy: managing in the next
and structural design. Manag. Sci. 32 (5), 554–571. generation o articial intelligence (AI). Cali. Manag. Rev. 61 (4), 43–65.
Darcy, A., Daniels, J., Salinger, D., Wicks, P., Robinson, A., 2021. Evidence o human- Jones, V.K., Hanus, M., Yan, C., Shade, M.Y., Boron, J.B., Bicudo, R.M., 2021. Reducing
level bonds established with a digital conversational agent: cross-sectional, loneliness among aging adults: the roles o personal voice assistants and
retrospective observational study. JMIR Formative Res. 5 (5), e27868. anthropomorphic interactions. Front. Public Health 9.
De Cicco, R., Silva, S.C., Alparone, F.R., 2020. Millennials’ attitude toward chatbots: an Kar, A.K., Kushwaha, A.K., 2021. Facilitators and barriers o articial intelligence
experimental study in a social relationship perspective. Int. J. Retail Distrib. Manag. adoption in business–insights rom opinions using big data analytics. In. Syst. Front.
48 (11), 1213–1233. 1–24.
De Gennaro, M., Krumhuber, E.G., Lucas, G., 2020. Eectiveness o an empathic chatbot Kasilingam, D.L., 2020. Understanding the attitude and intention to use smartphone
in combating adverse eects o social exclusion on mood. Front. Psychol. 3061. chatbots or shopping. Technol. Soc. 62, 101280.
Derlaga, V.J., Berg, J.H. (Eds.), 1987. Sel-disclosure: Theory, Research, and Therapy. Keestra, M., 2017. Metacognition and refection by interdisciplinary experts: insights
Springer Science & Business Media. rom cognitive science and philosophy. Issues Interdisc. Stud. 35, 121–169.
Donthu, N., Kumar, S., Mukherjee, D., Pandey, N., Lim, W.M., 2021a. How to conduct a Kessler, M.M., 1963. Bibliographic coupling between scientic papers. Am. Doc. 14 (1),
bibliometric analysis: an overview and guidelines. J. Bus. Res. 133, 285–296. 10–25.
Donthu, N., Kumar, S., Pattnaik, D., Lim, W.M., 2021b. A bibliometric retrospection o Kim, S., Choudhury, A., 2021. Exploring older adults’ perception and use o smart
marketing rom the lens o psychology: insights rom. Psychol. Mark. 38 (5), speaker-based voice assistants: a longitudinal study. Comput. Hum. Behav. 124,
834–865. 106914.
Dwivedi, Y.K., Kshetri, N., Hughes, L., Slade, E.L., Jeyaraj, A., Kar, A.K., Wright, R., Kowatsch, T., Nißen, M., Rüegger, D., Stieger, M., Flückiger, C., Allemand, M., von
2023. “So what i ChatGPT wrote it?” multidisciplinary perspectives on Wangenheim, F., 2018. The Impact o Interpersonal Closeness Cues in Text-Based
opportunities, challenges and implications o generative conversational AI or Healthcare Chatbots on Attachment Bond and the Desire to Continue Interacting: An
research, practice and policy. Int. J. In. Manag. 71, 102642. Experimental Design.
Eagly, A.H., Wood, W., 1999. The origins o sex dierences in human behavior: evolved Krämer, N.C., Eimler, S., Von Der Pütten, A., Payr, S., 2011. Theory o companions: what
dispositions versus social roles. Am. Psychol. 54, 408–423. can theoretical models contribute to applications and understanding o human-robot
Eisenhardt, K., 1989. Agency theory: a review and assessment. Acad. Manag. Rev. 14 (1), interaction? Appl. Arti. Intell. 25 (6), 474–502.
57–74. Kushwaha, A.K., Kumar, P., Kar, A.K., 2021. What impacts customer experience or B2B
Elyashar, A., Fire, M., Kagan, D., Elovici, Y., 2013, August. Homing socialbots: intrusion enterprises on using AI-enabled chatbots? Insights rom Big data analytics. Ind.
on a specic organization’s employee using socialbots. In: Proceedings o the 2013 Mark. Manag. 98, 207–221.
IEEE/ACM International Conerence on Advances in Social Networks Analysis and Laranjo, L., Dunn, A.G., Tong, H.L., Kocaballi, A.B., Chen, J., Bashir, R., Coiera, E., 2018.
Mining, pp. 1358–1365. Conversational agents in healthcare: a systematic review. J. Am. Med. Inorm. Assoc.
Epley, N., Waytz, A., Cacioppo, J.T., 2007. On seeing human: a three-actor theory o 25 (9), 1248–1258.
anthropomorphism. Psychol. Rev. 114 (4), 864. Lee, S., Lee, N., Sah, Y.J., 2020a. Perceiving a mind in a Chatbot: eect o mind
Far, S.B., Rad, A.I., 2018. Security Analysis o Big Data on Internet o Things. arXiv perception and social cues on co-presence, closeness, and intention to use. Int. J.
preprint arXiv:1808.09491. Hum. Comput. Interact. 36 (10), 930–940.
Feine, J., Gnewuch, U., Morana, S., Maedche, A., 2019. A taxonomy o social cues or Lee, Y.C., Yamashita, N., Huang, Y., 2020b. Designing a chatbot as a mediator or
conversational agents. Int. J. Hum. Comput. Stud. 132, 138–161. promoting deep sel-disclosure to a real mental health proessional. Pro. ACM
Floridi, L., 2008. Articial intelligence’s new rontier: articial companions and the Human-Comput. Interact. 4 (CSCW1), 1–27.
ourth revolution. Meta Philos. 39 (4–5), 651–655. Lee, Y.C., Yamashita, N., Huang, Y., Fu, W., 2020c. “I hear you, i eel you”: encouraging
Freedman, J.L., Fraser, S.C., 1966. Compliance without pressure: the oot-in-the-door deep sel-disclosure through a Chatbot. In: Proceedings o the 2020 CHI Conerence
technique. J. Pers. Soc. Psychol. 4 (2), 195. on Human Factors in Computing Systems, pp. 1–12.
Gama, S., Barata, G., Gonçalves, D., Prada, R., Paiva, A., 2011. SARA: social aective Lei, S.I., Shen, H., Ye, S., 2021. A comparison between chatbot and human service:
relational agent: a study on the role o empathy in articial social agents. In: customer perception and reuse intention. Int. J. Contemp. Hosp. Manag. 33 (11),
Aective Computing and Intelligent Interaction: 4th International Conerence, ACII 3977–3995.
2011, Memphis, TN, USA, October 9–12, 2011, Proceedings, Part I 4. Springer, Levinger, G., 1980. Toward the analysis o close relationships. J. Exp. Soc. Psychol. 16
Berlin Heidelberg, pp. 507–516. (6), 510–544.
18
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Leydesdor, L., Raols, I., 2011. Indicators o the interdisciplinarity o journals: diversity, Portacolone, E., Halpern, J., Luxenberg, J., Harrison, K.L., Covinsky, K.E., 2020. Ethical
centrality, and citations. J. In. Secur. 5 (1), 87–100. issues raised by the introduction o articial companions to older adults with
Lim, M.Y., 2012. Memory models or intelligent social companions. In: Human-computer cognitive impairment: a call or interdisciplinary collaborations. J. Alzheimers Dis.
Interaction: The Agency Perspective. Springer, Berlin, Heidelberg, pp. 241–262. 76 (2), 445–455.
Lim, W.M., Kumar, S., Verma, S., Chaturvedi, R., 2022. Alexa, what do we know about Portela, M., Granell-Canut, C., 2017, September. A new riend in our smartphone?
conversational commerce? Insights rom a systematic literature review. Psychol. Observing interactions with Chatbots in the search o emotional engagement. In:
Mark. 39 (6), 1129–1155. Proceedings o the XVIII International Conerence on Human Computer Interaction,
Lisetti, C.L., 1998. Aective Computing. pp. 1–7.
Lopatovska, I., Williams, H., 2018, March. Personication o the Amazon Alexa: BFF or a Possati, L.M., 2022. Psychoanalyzing articial intelligence: the case o Replika. AI & Soc.
mindless companion. In: Proceedings o the 2018 Conerence on Human Inormation 1–14.
Interaction and Retrieval, pp. 265–268. Potdevin, D., Clavel, C., Sabouret, N., 2021. Virtual intimacy in human-embodied
Loveys, K., Fricchione, G., Kolappa, K., Sagar, M., Broadbent, E., 2019. Reducing patient conversational agent interactions: the infuence o multimodality on its perception.
loneliness with articial agents: design insights rom evolutionary neuropsychiatry. J. Multimodal User Interaces 15, 25–43.
J. Med. Internet Res. 21 (7), e13664. Pradhan, A., Findlater, L., Lazar, A., 2019. "Phantom riend" or" just a box with
Lowry, R., 2015. September Meet the Lonely Japanese Men in Love With Virtual inormation" personication and ontological categorization o smart speaker-based
Girlriends. Time. Retrieved rom. http://time.com/3998563/virtual-love-japan/. voice assistants by older adults. In: Proceedings o the ACM on Human-Computer
Retrieved rom. Interaction, 3. CSCW, pp. 1–21.
MacLeod, M., 2018. What makes interdisciplinarity dicult? Some consequences o Premack, D., Woodru, G., 1978. Does the chimpanzee have a theory o mind? Behav.
domain specicity in interdisciplinary practice. Synthese 195 (2), 697–720. Brain Sci. 1 (4), 515–526.
Maroukhani, P., Asadi, S., Ghobakhloo, M., Jannesari, M.T., Ismail, W.K.W., 2022. How Preston, S.D., De Waal, F.B., 2002. Empathy: its ultimate and proximate bases. Behav.
do interactive voice assistants build brands’ loyalty? Technol. Forecast. Soc. Chang. Brain Sci. 25 (1), 1–20.
183, 121870. Provoost, S., Lau, H.M., Ruwaard, J., Riper, H., 2017. Embodied conversational agents in
McCroskey, L.L., McCroskey, J.C., Richmond, V.P., 2006. Analysis and improvement o clinical psychology: a scoping review. J. Med. Internet Res. 19 (5), e6553.
the measurement o interpersonal attraction and homophily. Commun. Q. 54 (1), Radicchi, F., Castellano, C., Cecconi, F., Loreto, V., Parisi, D., 2004. Dening and
1–31. identiying communities in networks. Proc. Natl. Acad. Sci. 101 (9), 2658–2663.
McGoldrick, P.J., Keeling, K.A., Beatty, S.F., 2008. A typology o roles or avatars in Radziwill, N., Benton, M., 2017. Evaluating quality o Chatbots and intelligent
online retailing. J. Mark. Manag. 24 (3–4), 433–461. conversational agents. Sotw. Qual. Pro. 19 (3), 25.
McGrath, J.E., 1991. Time, interaction, and perormance (TIP) a theory o groups. Small Ramadan, Z., Farah, F., M., and El Essrawi, L., 2021. From Amazon.com to Amazon.
Group Res. 22 (2), 147–174. Love: how Alexa is redening companionship and interdependence or people with
McLean, G., Osei-Frimpong, K., Barhorst, J., 2021. Alexa, do voice assistants infuence special needs. Psychol. Mark. 38 (4), 596–609.
consumer brand engagement?–examining the role o AI powered voice assistants in Ramadan, Z.B., 2021. “Alexaying” shoppers: the examination o Amazon’s captive
infuencing consumer brand engagement. J. Bus. Res. 124, 312–328. relationship strategy. J. Retail. Consum. Serv. 62, 102610.
Mensio, M., Rizzo, G., Morisio, M., 2018, April. The rise o emotion-aware conversational Rapp, A., Curti, L., Boldi, A., 2021. The human side o human-chatbot interaction: a
agents: threats in digital emotions. In: Companion Proceedings o the The Web systematic literature review o ten years o research on text-based chatbots. Int. J.
Conerence 2018, pp. 1541–1544. Hum. Comput. Stud. 151, 102630.
Mhatre, P., Gedam, V., Unnikrishnan, S., Verma, S., 2020. Circular economy in the built Rhee, C.E., Choi, J., 2020. Eects o personalization and social role in voice shopping: an
environment–literature review and theory development. J. Build. Eng. 101995. experimental study on product recommendation by a conversational voice agent.
Mick, D.G., Fournier, S., 1998. Paradoxes o technology: consumer cognizance, emotions, Comput. Hum. Behav. 109, 106359.
and coping strategies. J. Consum. Res. 25 (2), 123–143. Ryan, R.M., Deci, E.L., 2000. Sel-determination theory and the acilitation o intrinsic
Mori, M., 1970. Bukimi no tani [the uncanny valley]. Energy 7, 33–35. motivation, social development, and well-being. Am. Psychol. 55 (1), 68.
Mottet, T.P., Frymier, A.B., Beebe, S.A., 2006. Theorizing about instructional Sa, M.F., Al Sadrani, B., Mustaa, A., 2021. Virtual voice assistant applications
communication. In: Handbook o Instructional Communication: Rhetorical and improved expressive verbal abilities and social interactions in children with autism
Relational Perspectives, pp. 255–282. spectrum disorder: a single-subject experimental study. Int. J. Dev. Disabil. 1–13.
Moussawi, S., Benbunan-Fich, R., 2021. The eect o voice and humour on users’ Schuetzler, R.M., Grimes, G.M., Scott Giboney, J., 2020. The impact o chatbot
perceptions o personal intelligent agents. Behav. Inorm. Technol. 40 (15), conversational skill on engagement and perceived humanness. J. Manag. In. Syst. 37
1603–1626. (3), 875–900.
Mukherjee, D., Lim, W.M., Kumar, S., Donthu, N., 2022. Guidelines or advancing theory Shum, H.Y., He, X.D., Li, D., 2018. From Eliza to XiaoIce: challenges and opportunities
and practice through bibliometric research. J. Bus. Res. 148, 101–115. with social chatbots. Front. In. Technol. Electron. Eng. 19 (1), 10–26.
Murtarelli, G., Gregory, A., Romenti, S., 2021. A conversation-based perspective or Silverstone, R., Hirsch, E., 1992. Consuming Technologies: Media and Inormation in
shaping ethical human–machine interactions: the particular challenge o chatbots. Domestic Spaces. Routledge, London.
J. Bus. Res. 129, 927–935. Sinoo, C., van Der Pal, S., Henkemans, O.A.B., Keizer, A., Bierman, B.P., Looije, R.,
Nakanishi, H., Nakazawa, S., Ishida, T., Takanashi, K., Isbister, K., 2003, July. Can Neerincx, M.A., 2018. Friendship with a robot: children’s perception o similarity
sotware agents infuence human relations? Balance theory in agent-mediated between a robot’s physical and virtual embodiment that supports diabetes sel-
communities. In: Proceedings o the Second International Joint Conerence on management. Patient Educ. Couns. 101 (7), 1248–1255.
Autonomous Agents and Multiagent Systems, pp. 717–724. Siourti, C., Quintas, J., Ben-Moussa, M., Hanke, S., Nijdam, N.A., Konstantas, D., 2018.
Nass, C., Moon, Y., 2000. Machines and mindlessness: social responses to computers. In: Bi, Y., Kapoor, S. (Eds.), The CaMeLi Framework—A Multimodal Virtual
J. Soc. Issues 56 (1), 81–103. Companion or Older Adults.
Nass, C., Steuer, J., Tauber, E.R., 1994, April. Computers are social actors. In: Skjuve, M., Følstad, A., Fostervold, K.I., Brandtzaeg, P.B., 2021. My chatbot companion-a
Proceedings o the SIGCHI Conerence on Human Factors in Computing Systems, study o human-chatbot relationships. Int. J. Hum. Comput. Stud. 149, 102601.
pp. 72–78. Smith, B. (Ed.), 2003. John Searle. Cambridge University Press, Cambridge.
Odekerken-Schröder, G., Mele, C., Russo-Spena, T., Mahr, D., Ruggiero, A., 2020. Sridevi, G.M., Suganthi, S.K., 2022. AI based suitability measurement and prediction
Mitigating loneliness with companion robots in the COVID-19 pandemic and between job description and job seeker proles. Int. J. In. Manag. Data Insights 2
beyond: an integrative ramework and research agenda. J. Serv. Manag. 31 (6), (2), 100109.
1149–1162. Suwono, L.V., Sihombing, S.O., 2016. Factors aecting customer loyalty o tness
Omrani, N., Rivieccio, G., Fiore, U., Schiavone, F., Agreda, S.G., 2022. To trust or not to centers: an empirical study. JDM (Jurnal Dinamika Manajemen) 7 (1), 45–55.
trust? An assessment o trust in AI-based systems: concerns, ethics and contexts. Sweller, J., 1988. Cognitive load during problem solving: eects on learning. Cogn. Sci.
Technol. Forecast. Soc. Chang. 181, 121763. 12, 257–285.
Orabi, M., Mouheb, D., Al Aghbari, Z., Kamel, I., 2020. Detection o bots in social media: Ta, V., Grith, C., Boateld, C., Wang, X., Civitello, M., Bader, H., Loggarakis, A., 2020.
a systematic review. In. Process. Manag. 57 (4), 102250. User experiences o social support rom companion chatbots in everyday contexts:
Payr, S., 2011. Social engagement with robots and agents: introduction. Appl. Arti. thematic analysis. J. Med. Internet Res. 22 (3), e16235.
Intell. 25 (6), 441–444. Takayanagi, K., Kirita, T., Shibata, T., Shibata, T., Shibata, T., 2014. Comparison o
Pesty, S., Duhaut, D., 2011, December. Articial companion: building a impacting verbal and emotional responses o elderly people with mild/moderate dementia and
relation. In: 2011 IEEE International Conerence on Robotics and Biomimetics. IEEE, those with severe dementia in responses to seal robot, PARO. Front. Aging Neurosci.
pp. 2902–2907. 6, 257-257. Retrieved 3 19, 2023.
Petronio, S., Durham, W., 2008. Understanding and applying communication privacy Tassiello, V., Tillotson, J.S., Rome, A.S., 2021. “Alexa, order me a pizza!”: the mediating
management theory. In: Engaging theories in interpersonal communication, role o psychological power in the consumer–voice assistant interaction. Psychol.
pp. 309–322. Mark. 38 (7), 1069–1080.
Pilkington, A., Catherine, L.H., 1999. Is production and operations management a Thorne, S., 2020. Hey Siri, tell me a story: digital storytelling and AI authorship.
discipline? A citation/co-citation study. Int. J. Oper. Prod. Manag. 19 (1), 7–20. Convergence 26 (4), 808–823.
Pitardi, V., Marriott, H.R., 2021. Alexa, she’s not human but… unveiling the drivers o Tsai, W.H.S., Liu, Y., Chuan, C.H., 2021. How chatbots’ social presence communication
consumers’ trust in voice-based articial intelligence. Psychol. Mark. 38 (4), enhances consumer engagement: the mediating role o parasocial interaction and
626–642. dialogue. J. Res. Interact. Mark. 15 (3), 460–482.
Porra, J., Lacity, M., Parks, M.S., 2020. “Can computer based human-likeness endanger Tsiourti, C., Quintas, J., Ben-Moussa, M., Hanke, S., Nijdam, N.A., Konstantas, D., 2016,
humanness?”–a philosophical and ethical perspective on digital assistants expressing September. The CaMeLi ramework—a multimodal virtual companion or older
eelings they can’t have. In. Syst. Front. 22 (3), 533–547. adults. In: Proceedings o SAI Intelligent Systems Conerence. Springer, Cham,
pp. 196–217.
19
R. Chaturvedi et al. Technological Forecasting & Social Change 193 (2023) 122634
Turunen, Markku, Hakulinen, Jaakko, Ståhl, Olov, Gambäck, Björn, Hansen, Preben, Sanjeev Verma is presently working as Proessor (Marketing) at National Institute o
Rodríguez, Mari C., Gancedo, Raúl Santos, de La Cámara, Cameron, Smith, Daniel Industrial Engineering (NITIE) Mumbai, India. Dr. Verma is an active researcher and has
Charlton, Cavazza, Marc, 2011. Multimodal and mobile conversational health and authored/co-authored more than 70 publications in reereed International/National
tness companions. Comput. Speech Lang. 25 (2), 192–209. journals and conerence proceedings and Authored one book. His papers have been pub-
Van Eck, N.J., Waltman, L., 2011. Text Mining and Visualization Using VOSviewer. arXiv lished in various international journals o Repute such as Psychology and Marketing,
preprint arXiv:1109.2058. Journal o Interactive Marketing, Tourism Review, Government Inormation Quarterly,
Vázquez-Cano, E., Mengual-Andrés, S., López-Meneses, E., 2021. Chatbot to improve Journal o Marketing Communications, Journal o Internet Commerce, Journal o
learning punctuation in Spanish and to enhance open and fexible learning Modeling Management, Journal o Global Marketing, International Journal o Marketing
environments. Int. J. Educ. Technol. High. Educ. 18 (1), 1–20. and Philanthropy etc.
Verma, S., 2022. Sentiment analysis o public services or smart society: literature review
and uture research directions. Gov. In. Q. 101708.
Ronnie Das is an Associate Proessor o Digital and Data Driven Marketing at Audencia
Verma, S., Yadav, N., 2021. Past, present, and uture o electronic word o mouth
Business School, France. He is also Head o the Marketing Department. Prior to joining
(EWOM). J. Interact. Mark. 53, 111–128.
Audenica, Ronnie has spent over a decade in UK higher education researching emerging
Verma, S., Sharma, R., Deb, S., Maitra, D., 2021. Articial intelligence in marketing:
technologies and application o machine learning in understanding marketing 4.0 and
systematic review and uture research direction. Int. J. In. Manage. Data Insights 1
transormative consumer behaviour. Ronnie's research, with Newcastle University Urban
(1), 100002.
Observatory, on Big Data driven insight into citizen behaviour during COVID crisis was
Walther, J.B., Burgoon, J.K., 1992. Relational communication in computer-mediated
published by the World Economic Forum and London School o Economics Impact Blog.
interaction. Hum. Commun. Res. 19 (1), 50–88.
Ronnie has also chaired prestigious research workstreams and research hub (SuperGen
Wang, L., Wang, D., Tian, F., Peng, Z., Fan, X., Zhang, Z., Wang, H., 2021. Cass: towards
Energy Network) committees unded by the UKRI (EPSRC). He has also successully
building a social-support chatbot or online health community. Proc. ACM Human-
developed and delivered projects unded and supported by ING Economics. Ronnie has
Comput Interact 5 (CSCW1), 1–31.
urther own applied innovations challenges organized by the European Commission
Whittemore, R., Chao, A., Jang, M., Minges, K.E., Park, C., 2014. Methods or knowledge
including EU Datathon 2020 and European Innovation Sprint 2020. He was also invited to
synthesis: An overview. Heart & Lung 43 (5), 453–461.
talk at the European Central Bank.
Wilson-Nash, C., Goode, A., Currie, A., 2020. Introducing the socialbot: a novel
touchpoint along the young adult customer journey. Eur. J. Mark. 54 (10),
2621–2643. Yogesh K Dwivedi is a Proessor o Digital Marketing and Innovation and Founding Di-
Wood, L.J., Zaraki, A., Robins, B., Dautenhahn, K., 2019. Developing Kaspar: a humanoid rector o the Emerging Markets Research Centre (EMaRC) at the School o Management,
robot or children with autism. Int. J. Soc. Robot. 1–18. Retrieved 3 19, 2023. Swansea University, Wales, UK. In addition, he holds a Distinguished Research Proes-
Youn, S., Jin, S.V., 2021. “In AI we trust?” The eects o parasocial interaction and sorship at the Symbiosis Institute o Business Management (SIBM), Pune, India. Proessor
technopian versus luddite ideological views on chatbot-based customer relationship Dwivedi is also currently leading the International Journal o Inormation Management as its
management in the emerging “eeling economy”. Comput. Hum. Behav. 119, Editor-in-Chie. His research interests are at the interace o Inormation Systems (IS) and
106721. Marketing, ocusing on issues related to consumer adoption and diusion o emerging
Zhou, L., Gao, J., Li, D., Shum, H.Y., 2020. The design and implementation o xiaoice, an digital innovations, digital government, and digital and social media marketing particu-
empathetic social chatbot. Comput. Linguist. 46 (1), 53–93. larly in the context o emerging markets. Proessor Dwivedi has published more than 500
Zupic, I., Cater, T., 2015. Bibliometric methods in management and organization. Organ. articles in a range o leading academic journals and conerences that are widely cited
Res. Methods 18 (3), 429–472.2w2ww. (more than 41 thousand times as per Google Scholar). He has been named on the annual
Highly Cited Researchers™ 2020 and 2021 lists rom Clarivate Analytics. Proessor Dwi-
vedi is an Associate Editor o the Journal o Business Research, European Journal o Mar-
Rijul Chaturvedi is a Fellow at National Institute o Industrial Engineering (NITIE),
keting, Government Inormation Quarterly and International Journal o Electronic Government
Mumbai, India. His research interest includes emotional AI in marketing, Conversational
Research, and Senior Editor o the Journal o Electronic Commerce Research.
commerce, and experiential marketing.
20
You might also like
- 5 Review - Allouch等 - 2021 - HCISE - Conversational Agents - goals Technologies Vision and ChallengesNo ratings yet5 Review - Allouch等 - 2021 - HCISE - Conversational Agents - goals Technologies Vision and Challenges48 pages
- Psychological Insights Into The Research and Practice of Embodied Conversational Agents Chatbots and Social Assistive Robots A Systematic Meta-RevieNo ratings yetPsychological Insights Into The Research and Practice of Embodied Conversational Agents Chatbots and Social Assistive Robots A Systematic Meta-Revie42 pages
- Defining AI Companions A Research Agenda-From ArtiNo ratings yetDefining AI Companions A Research Agenda-From Arti45 pages
- Servant by Default How Humans Perceive Their Relationship WithNo ratings yetServant by Default How Humans Perceive Their Relationship With23 pages
- 4 - (2019) Guzman - Artificial Intelligence and CommunicationNo ratings yet4 - (2019) Guzman - Artificial Intelligence and Communication17 pages
- UX Research On Conversational Human-AI Interaction: A Literature Review of The ACM Digital LibraryNo ratings yetUX Research On Conversational Human-AI Interaction: A Literature Review of The ACM Digital Library24 pages
- From Eliza To XiaoIce Challenges and Opportunities With Social ChatbotsNo ratings yetFrom Eliza To XiaoIce Challenges and Opportunities With Social Chatbots20 pages
- Compassionate AI Design, Governance, and UseNo ratings yetCompassionate AI Design, Governance, and Use7 pages
- Future Directions For Chatbot Research: An Interdisciplinary Research AgendaNo ratings yetFuture Directions For Chatbot Research: An Interdisciplinary Research Agenda28 pages
- nerative AI and human-robot interaction implications and future agenda for business, society and ethics生成式人工智能和人机交互：对商业、社会和道德的影响和未来意图No ratings yetnerative AI and human-robot interaction implications and future agenda for business, society and ethics生成式人工智能和人机交互：对商业、社会和道德的影响和未来意图14 pages
- My Chatbot Companion - A Study of Human-Chatbot RelationshipsNo ratings yetMy Chatbot Companion - A Study of Human-Chatbot Relationships14 pages
- Guzman Lewis AIand Communication NMS2019No ratings yetGuzman Lewis AIand Communication NMS201918 pages
- The Evolution, Applications, and Ethical Implications of ChatGPT A Study of AI Powered Conversational AgentsNo ratings yetThe Evolution, Applications, and Ethical Implications of ChatGPT A Study of AI Powered Conversational Agents4 pages
- Done - When Chatbots Fail Exploring User Coping Following A Chatbots-Induced Service FailureNo ratings yetDone - When Chatbots Fail Exploring User Coping Following A Chatbots-Induced Service Failure21 pages
- From HAL To GenAI, Optimizing Chatbot Impacts With CARENo ratings yetFrom HAL To GenAI, Optimizing Chatbot Impacts With CARE12 pages
- Improving Accessibility and Readability of Survey Reports in Digital Health Platforms Using Conversational AINo ratings yetImproving Accessibility and Readability of Survey Reports in Digital Health Platforms Using Conversational AI3 pages
- Effects of Gender and Relationship Type On The ResponseNo ratings yetEffects of Gender and Relationship Type On The Response5 pages
- Evolving Trends Limitations and Ethical Considerations in AI-driven Conversational Interfaces Assessing ChatGPT S Impact On Healthcare Financial SNo ratings yetEvolving Trends Limitations and Ethical Considerations in AI-driven Conversational Interfaces Assessing ChatGPT S Impact On Healthcare Financial S21 pages
- Establishing Credibility in AI Chatbots The Importance of Customization, Communication Competency and User SatisfactionNo ratings yetEstablishing Credibility in AI Chatbots The Importance of Customization, Communication Competency and User Satisfaction18 pages
- Human-Agent Interaction and Human Dependency: Possible New Approaches For Old ChallengesNo ratings yetHuman-Agent Interaction and Human Dependency: Possible New Approaches For Old Challenges10 pages
- Institutional Activitiy Paper Nguvii, Asaf, AbhiNo ratings yetInstitutional Activitiy Paper Nguvii, Asaf, Abhi4 pages
- Communicating Arti Cial Intelligence (AI) - Theory, Research, and PracticeNo ratings yetCommunicating Arti Cial Intelligence (AI) - Theory, Research, and Practice5 pages
- From Eliza To XiaoIce Challenges and Opportunities With Social ChatbotNo ratings yetFrom Eliza To XiaoIce Challenges and Opportunities With Social Chatbot17 pages
- Chatbot As Social Companionshow People Perceive Consciousness, Huamn Likeness, and Social Health Benefits in MachinesNo ratings yetChatbot As Social Companionshow People Perceive Consciousness, Huamn Likeness, and Social Health Benefits in Machines14 pages
- CHI18 SIG Chatbots+for+social+good Authors+versionNo ratings yetCHI18 SIG Chatbots+for+social+good Authors+version5 pages
- Enhancing Communication in Human AI InteractionNo ratings yetEnhancing Communication in Human AI Interaction7 pages
- Journal of Business Research: Grazia Murtarelli, Anne Gregory, Stefania RomentiNo ratings yetJournal of Business Research: Grazia Murtarelli, Anne Gregory, Stefania Romenti9 pages
- A Brief Overview of Chatgpt: The History, Status Quo and Potential Future Development. Ieee/Caa Journal of Automatica SinicaNo ratings yetA Brief Overview of Chatgpt: The History, Status Quo and Potential Future Development. Ieee/Caa Journal of Automatica Sinica2 pages
- SurveyonIntelligentChatbots - State of The ArtandFutureResearchDirectionsNo ratings yetSurveyonIntelligentChatbots - State of The ArtandFutureResearchDirections11 pages
- SurveyonIntelligentChatbots - State of The ArtandFutureResearchDirectionsNo ratings yetSurveyonIntelligentChatbots - State of The ArtandFutureResearchDirections11 pages
- Chatgpt: Future Directions and Open Possibilities: Editorial ArticleNo ratings yetChatgpt: Future Directions and Open Possibilities: Editorial Article2 pages
- Chatgpt: Future Directions and Open Possibilities: January 2023No ratings yetChatgpt: Future Directions and Open Possibilities: January 20233 pages
- Survey of Various AI Chatbots Based On Technology Used: Abstract-Artificial Intelligence Is The Ability To Make100% (1)Survey of Various AI Chatbots Based On Technology Used: Abstract-Artificial Intelligence Is The Ability To Make6 pages
- Ch5 - Items of FS - CA Inter Audit - Ankush Chirimar Classes100% (1)Ch5 - Items of FS - CA Inter Audit - Ankush Chirimar Classes16 pages
- Rammed Earth and Bamboo Construction-PresentationNo ratings yetRammed Earth and Bamboo Construction-Presentation36 pages
- Homework: 1 CAP315: Object Oriented Analysis and DesignNo ratings yetHomework: 1 CAP315: Object Oriented Analysis and Design14 pages
- Using Role-Based Permissions: Public Document Version: Q4 2019 - 2020-02-01No ratings yetUsing Role-Based Permissions: Public Document Version: Q4 2019 - 2020-02-01170 pages
- Mijoni V Zambia Publishing Company LTD (Appeal 10 of 1986) 1987 ZMSC 35 (27 July 1987)No ratings yetMijoni V Zambia Publishing Company LTD (Appeal 10 of 1986) 1987 ZMSC 35 (27 July 1987)15 pages
- Auger Boring: Name: Aabhash Bhattarai Roll No.: 07100% (1)Auger Boring: Name: Aabhash Bhattarai Roll No.: 076 pages
- Fossil Group: Process/Scenario: MM - Business - Partner - Creation - (Supplier)No ratings yetFossil Group: Process/Scenario: MM - Business - Partner - Creation - (Supplier)39 pages
- Information Technology Class X Total Marks: 100 (Theory - 50 + Practical-50) Theory Unit No. Unit Name MarkNo ratings yetInformation Technology Class X Total Marks: 100 (Theory - 50 + Practical-50) Theory Unit No. Unit Name Mark2 pages