Background:

AI is emerging as a key component of the digital world in which university

students are increasingly being transformed in terms of how they access, process,
and engage with information within their education. These range from AI-powered
search engines and chatbots to personalized learning platforms; as a result, students
are turning to intelligent systems more and more to assist them in academia. While
AI provides convenience and efficiency, its increasing presence also begs the
question of how it may impede cognitive characteristics like curiosity and
information-seeking behavior—two fundamental drivers of deep learning and
intellectual engagement.
Curiosity, which is typically defined as motivation or desire for the acquisition of
new knowledge or experience, is a critical component required for academic
success and lifelong learning. Information Seeking Behavior is the common name
for the ways persons look for and acquire information to meet their knowledge
needs. As A.I. gets better at predicting users’ questions and providing ready-made
answers, it could reinforce or undermine these cognitive motivations. Reframing
educational institutions in this direction can help them deliver content while also
enabling curiosity and active engagement with what is being taught.
Problem statement:
Use of AI may have benefits in improving access to information for students, but
the increasing reliance on AI tools may inhibit students’ natural curiosity and
lessen the energy they spend acquiring knowledge on their own. Because you can
find “the answer” at click of a mouse (without, as the saying goes, recording “the
process”), the questioning, exploring, and critically engagement with information
may simply be skipped. On the other hand, AI may inspire students to burn more
fuel on the flame of curiosity by leading students to new ideas and personalized
recommendations that help motivate them to dig more. However, empirical
research investigating this dual impact—particularly in university settings—is still
limited.
Artificial intelligence offers numerous benefits that have transformed the
educational experience for university students. One of the most significant
advantages is efficiency and speed—AI tools can process and deliver vast amounts
of information within seconds, enabling students to complete research tasks more
quickly. Additionally, AI-powered platforms offer personalized learning
experiences, adapting to individual learning styles, needs, and progress, which can
enhance student engagement and academic performance. Moreover, AI increases
accessibility, particularly for students with disabilities or those facing educational
barriers, by providing tools like text-to-speech, real-time translation, and adaptive
assessments. These benefits collectively create a learning environment that is more
responsive, inclusive, and tailored to individual academic journeys.
However, the increasing reliance on AI in academic settings also presents several
drawbacks. One major concern is the potential for cognitive passivity, were overly
dependent on AI-generated responses, reducing them to actively seek information
or think critically. The convenience of instant answers might lead to reduced
problem-solving and analytical skills, as students may bypass the deeper cognitive
processes involved in inquiry and research. Furthermore, there is a risk of
diminished curiosity, as the natural drive to explore, question, and discover may be
weakened when answers are readily available without the need for effortful search
or reflection. These drawbacks suggest that while AI can support learning, it may
also unintentionally hinder the development of essential intellectual traits if not
used thoughtfully.

Literature review:
Artificial Intelligence:
Over the past 50 years, a variety of technologies have been created to enable
machines to perform tasks that have traditionally required human intellect, such as
sensing, reasoning, learning, and interacting. These technologies are together
referred to as artificial intelligence (AI). However, recent breakthroughs in
generative AI (GenAI), notably models like ChatGPT, have brought unprecedented
attention to AI’s revolutionary potential across numerous industries. Generative AI
is focused on producing new material, such as writing, images, and codes, by
utilizing deep learning models. This contrasts with predictive (pre-generative) AI,
which involves predictions and decision making through a variety of machine
learning and modeling techniques (Dai, 2023). This distinction is critical to
comprehend the scope of AI applications in education. A subfield of computer
science known as artificial intelligence (AI) replicates or simulates human
intelligence via the application of algorithms and machine learning techniques
(Helm et al. 2020).
Types of AI:
Artificial Superintelligence, Narrow AI, and General AI are the three categories of
AI. The most prevalent and developed type of AI to date is narrow AI. Utilizing
machine learning tools, such as image and facial recognition and Siri/Alexa, it is
very goal-oriented and works toward a single objective or task. Artificial
intelligence (AI) that is comparable to human intelligence (e.g., AI that can detect
the wants and emotions of other intelligent entities) is known as general AI (or
deep AI). Thirdly Artificial Superintelligence refers to AI that is more powerful
than humans (think of a sci-fi movie were dominates human society on all fronts).
A future full of intriguing possibilities is promised by the incorporation of AI into
education. Its capacity to enhance learning outcomes is among its most important
advantages. Precise education based on each student's strengths and shortcomings
help them understand subjects more fully. Teachers can provide focused
interventions and make sure all students attain their full potential since AI can
identify students who might need extra help. AI has the potential to greatly
increase classroom productivity. According to Gökçe Arslan et al. (2024), teachers
can allocate more chance to student development and generating engaging learning
experiences by automating administrative activities such as grading and attendance
tracking. AI can also improve resource allocation, guaranteeing that all students
have easy access to important learning resources and tools.
With the emergence of virtual assistants like Siri and Google Assistant, as well as
numerous other AI enabled applications in a wide range of industries like
healthcare, automotive, education, social media, entertainment, and robotics,
artificial intelligence (AI) is growing up alongside a generation of students in a
digital world that is changing quickly (Druga et al., 2018). According to Wang
(2020), artificial intelligence (AI) is the science and engineering of solving
problems using cutting-edge technologies like machine learning and neural
networks. It stands for the integration of STEM fields—science, technology,
engineering, and mathematics—which is important in today's technologically
advanced society.
Artificial Intelligence (AI) is the simulation of human intelligence processes by
machines, especially computer systems. These processes are learning, reasoning,
problem-solving, perception, and language understanding. AI systems are
computer systems that must be able to carry out duties that, normally, require
human intelligence for example visual perception, speech recognition, decision-
making and translation between languages (Russell & Norvig, 2016). AI, proving
intuitive and powerful solutions in sectors like healthcare, finance, education, and
transportation has had great influence over the years, augmenting our human
capabilities. AI is a broad field that includes areas such as machine learning (ML),
natural language processing (NLP), robotics, and expert systems. One major sub-
area of AI is machine learning, algorithms that allow systems to learn such that
they make predictions or decisions based on input data (Goodfellow, Bengio, &
Courville, 2016).
Ahmed and Lyu (2023) introduce the Artificial Intelligence-Enabled Intelligent
Assistant (AIIA), a framework leveraging AI and natural language processing to
facilitate personalized, adaptive learning in higher education. By providing
personalized learning journeys, answering questions, and creating quizzes, AIIA
streamlines cognitive function and keeps students learning. Incorporated with
learning management systems, it offers an interactive setting that supports
curiosity and efficient information-gathering. The paper discusses AIIA’s promise
for revolutionizing virtual teaching assistants, enhancing learning outcomes,
challenges and opportunities, and future research avenues.
In contrast, natural language processing is specifically concerned with the
interaction between computers and human languages, allowing computers to
understand, interpret, and produce human language (Jurafsky & Martin, 2021).
Robotics is the field of designing and constructing robots capable of
independently performing tasks, frequently using AI methods to enhance
efficiency and versatility (Siciliano & Khatib, 2016). The availability of large,
complex datasets has been one of the biggest contributors to AI advancement,
providing ample information to train the new models. By analyzing existing data,
these models can recognize patterns and make predictions, which is proving to be
an asset to the decision-making process (Minsky, 1967). Yet as its capabilities
have grown, AI has raised questions of ethics, transparency, and accountability.
Concerns like algorithmic bias, the risk of job displacement, and the ethical use of
AI in autonomous weapons systems are now hot topics for researchers and
policymakers alike (Binns, 2018; O'Neil, 2016).
Curiosity:
Curiosity is one of the most elementary psychical characteristics to satisfy quest
for fresh details, examine surroundings and learn about world. It has innate value
as it is based on intrinsic motivation to learn, explore and resolve uncertainty
(Loewenstein, 1994). People question, explore new experiences and engage in
behaviors that aim to fill cognitive gaps through curiosity. Because curiosity is an
essential human capacity, fundamental to cognitive growth (Silvia, 2008), the role
of effective questions in developing and enriching curiosity can hardly be
underestimated. In psychology, curiosity is generally divided into two main types
diversive curiosity and specific curiosity.
Diversive curiosity is the broad motivation to experience novel and excitatory
stimuli, based on novelty and sensory stimulation. Specific curiosity, on the other
hand, is more directed, where knowledge is sought to fill specific voids in
knowledge or to resolve a particular problem (Loewenstein, 1994). Both types of
curiosity cooperate to create an ongoing cycle of exploration and learning, within
both informal general life circumstances and more formal environments like
school and work. This relates to more than our individual behavior; the effect of
curiosity also contributes to social dynamics and group interaction. Because it
motivates individuals to ask about others' experiences, perspectives, and emotions,
curiosity can facilitate social learning and communication (Kashdan & Steger,
2007)
Curiosity in the workplace is considered a great trait because it encourages
innovation, adaptability, and knowledge sharing (Searle, 2019). Such employees
are more likely to try creative solutions inspired by their curiosity, which among
other things prepares their minds for collaboration (Murayama et al., 2013).
Scientific research has also examined the neurological and biological bases of
curiosity. Research has shown that the brain regions that activate during curiosity
are those related to reward and motivation, including the caudate nucleus and the
anterior cingulate cortex (Gruber et al., 2014). These studies imply that curiosity is
more than a mental and emotional phenomenon; it is a biological mechanism that
shapes the behavior and choices we make.
In addition, using a more numerous quantities of curiosity are linked with
cognitive flexibility and resilience because an open-mindedness and willingness to
learn from adversity (Reio, 2016). While curiosity has many positive attributes, it
can also create challenges. Indeed, too much curiosity, especially for certain types
of information, can lead to distraction, information overload, or even anxiety (Van
Hoorn et al., 2016). People with high curiosity levels, for example, might become
too buzzed about seeking answers, to the detriment of responsibilities or
relationships. Consequently, grasping how attention balances curiosity with focus
is an area that warrants further investigation.

Information Seeking Behavior:

The process through which individuals actively seek and acquire information in
response to a need or gap in their knowledge. This behavior is crucial for
decision-making, problem-solving, and learning in a wide range of fields,
including education, healthcare, business, and personal development (Wilson,
1999). Information seeking involves a process, not just searching for data — a set
of actions that includes realizing that you need information, seeking it, evaluating
it, and using it (Kuhlthau, 2004). It is fundamental for human cognition and
integral for adapting to new environments and developing and achieving better
individual and social results. Kulthau's Information Search Process (ISP) model is
one of the early models of its kind, describing the stages of emotional and
cognitive activity that a person goes through while searching for information.
The research process, phases include initiation, selection, exploration, formulation,
collection and presentation, was described as a fluid cycle by Kuhlthau (2004).
During these phases, there may well be a sense of doubt, confusion, and
frustration that the information being sought not to mention the process for
achieving it if the exposed information appears simple or ambiguous. On the
contrary, satisfactory information searching results the end of uncertainty and
fulfilling its information need (Kuhlthau, 2004).
Some Factors Affecting Information Seeking Behavior For example, cognitive
styles (e.g., a preference for analytical versus intuitive approaches) may influence
how individuals search for and process information (Eppler & Mengis, 2004).
Moreover, the digitalization of tools and resources in the recent past has changed
the technological landscape of information seeking behavior. It has often made
information more accessible but has also given rise to issues like information
overload and misinformation (Bawden & Robinson, 2009).
Likewise in academia, students may engage in a shape of information seeking
behavior to achieve learning goals or to fulfil certain type of research tasks (Julien
& Barker, 2009). The effectiveness and efficiency of the search process is also
dependent on the perceived value of the information and confidence in the
searcher’s ability to discover the right information.
For a strong start in today's day and age, one has trouble understanding the quality,
relevance, and credence of the information he/she sees. The underlying aspect of
information seeking behavior is the information needs of the individual. The
motivation to learn can vary based on the urgency and importance of the
information need. For instance, in high-stakes settings like medical care, people
might perform intensive and targeted information seeking to make knowledgeable
decisions about their health (Johnson, 2003).
Social and contextually focused factors are similarly important in influencing
information seeking behavior. Individuals tend to rely on social networks, peer
influence, or certain institutional support systems in any information search.
Individuals often depend on peers, co-workers, friends, or online communities
during uncertain or ambiguous situations for recommendations and advice (Fidel,
1993).
Cultural and environmental factors also contribute to the types of information
requested as well as strategies for finding information. In some cultures, it might
be more of a group activity, whereas, in others, it might be more of an individual
one. In addition, seeking information is not always a linear or intentional process.
It may be cyclical and flexible, as people may loop back to earlier steps or change
their approach due to new knowledge or changing needs (Savolainen, 2008). This
flexibility is important in living environments where people do not know how they
may use, what they will be exposed to, and what other news sources will be
available.
Traditionally, the study of human information-seeking behavior has focused on
the instrumental value of knowledge, where information plays a role in decision-
making and goal accomplishment. Yet Golman and Loewenstein (2015) propose a
bigger and more finely grained framework combining several psychological
motives for seeking or avoiding information. Their argument weaves together
elements of curiosity, information gap, and attention dynamics, leading to the
suggestion that knowledge carries emotional valence, being intrinsically rewarding
or aversive when attained or missing.
Their model is based on information-gap theory of curiosity (Loewenstein, 1994) if
curiosity increases when the person feels the discrepancy between their current
knowledge and their target knowledge. Golman and Loewenstein (2015) extend
this theory by adding the utility of knowledge more broadly, beyond its
instrumental use. For example, they claim that people occasionally search for
information just because they enjoy thinking about certain information while
intentionally avoiding information that can create emotional discomfort, which is
in line with “ostrich effect” (Karlsson et al., 2009).
In recent years, one subfield of machine learning, deep learning, has drawn a lot
of attention. Neural networks include multiple layers of them, which can learn
from vast amounts of data and execute complex tasks like image and speech
recognition (LeCun, Bengio, & Hinton, 2015).
The Synergistic Impact
In any case, the relationship between curiosity, information seeking behavior, and
artificial intelligence is complex and likely synergistic. Where curiosity drives
information seeking, AI's ability to connect users with content to explore makes
every search process more efficient and tailored. AI-based technologies enable
knowledge acquisition in faster, more efficient ways and may cultivate a person’s
curiosity by introducing new, interesting, and applicable information in
compelling formats. But curiosity may also be influenced and shaped by these
same technologies as algorithmic recommendations (AI that generates suggestions)
can help facilitate information journeys by introducing the individuals to
information, it can constrain the exploration. Thus, potentially for guidelines and
magnifying the person's interest, leading to decreased curiosity over time (Zeng et
al., 2019).
Additionally, AI can have an influence on how its creativity is fostered. For
example, if we think about the utilization of AI in an educational context, consider
how AI-powered adaptive learning platforms adjust the content delivery process
depending on diverse learning styles and curiosity levels in learners, ensuring a
unique educational experience that encourages behavioral engagement, typically in
forms of deeper engagement (Heffernan & Heffernan, 2014). Tailoring experiences
based on the learner's needs allows these platforms to set the stage for a relevant
environment that fosters curiosity, creating desire and direction for the search and
resolution of gaps that arise.
As with many types of relationship, the relationship between curiosity and
information-seeking behavior is one of symbiosis, with each influencing the other.
Research shows that more curious people tend to seek out information, but the act
of seeking out information can also increase curiosity, leading to a positive
feedback loop in which each one reinforces the other. When someone finds
conflicting information, sparking a need to know More (Wilson, 1999), they may
pursue it over and over in a nearly cyclical manner (i.e. curiosity-driven behavior).
However, the relationship between curiosity and information seeking is a feedback
loop that is critical to encouraging lifelong learning and intellectual development
(Loewenstein, 1994).
AI's role in supporting information seeking is especially clear within educational
and professional contexts. AI tools such as intelligent tutoring systems and
adaptive learning platforms aim to improve access to information and create
customized learning experiences (Heffernan & Heffernan, 2014). These
technologies can evaluate a learner's needs and interests, providing content
compatible with their level of knowledge and curiosity. Utilizing AI-Powered
Feedback One of the most valuable aspects of AI is its influence on active learning
experiences, as research shows that AI feedback systems (when used
appropriately) can boost engagement levels amongst students through
contextualizing information and providing real-time feedback to navigate complex
concepts in a personalized manner (Gruber et al., 2014).
But the rising use of artificial intelligence in information-seeking activity comes
with its own challenges. Social media algorithms and AI-driven content delivery
systems can contribute to information dysphoria, due to the volume of information
that users are inundated with. This phenomenon is even more acute in the case of
digital platforms where the amount of user-addled content is so high that
information is personalized to the point that users might feel bombarded (Bawden
& Robinson, 2009).
Moreover, although AI can foster curiosity through new content, it may also
unintentionally restrict the range of information presented to users since
algorithms tend to deploy user preferences rather than exposing them to less
preferable or challenging material (Pariser, 2011). Moreover, AI could clearly
shape the way that curiosity is cultivated and expressed. AI can create a
personalized learning experience for students, allowing them to explore topics of
interest while simultaneously adjusting the difficulty level to match a student’s
capabilities in a way that ensures that they remain engaged and inspired.
The empirical evidence from various studies suggests that AI platforms can
motivate a student to explore beyond existing plateaus by continuous immediate
feedback and conceptual emphasis on domain areas for further exploration
(Murayama et al., 2013). AI can track student performance and modify the learning
experience accordingly, so more personalized pathways can enhance curiosity and
engagement with the subject at hand. This is where AI also plays a key role in
overcoming the challenges that information seeking can bring. AI systems save
significant time by aggregating and analyzing information from multiple
traditional sources, even if the sources are often inaccurate (Russell & Norvig,
2016).
This efficiency can lower cognitive load, freeing users to spend more time
engaging with the information rather than searching for it. Given that the increased
efficiency of information retrieval through AI may also undermine the motivation
to engage in deeper, more exploratory forms of information seeking—over reliance
on AI-generated information (Zeng et al., 2019)—the ability to swap these two
modes of seeking may become less common in an age where excessive reliance on
AI is encouraged.
Wu and Chen (2016) aimed to investigate how levels of curiosity and uncertainty
influence how children search for information in a virtual world and whether
children take a shortcut "survey" approach or a detailed "route" approach.
Motivation had a significant influence on search strategy compared to other factors
in their study, it increased the likelihood of using a central, detail-rich route among
children with positive motivation, while those with uncertainty or negative emotion
favor a heuristic survey strategy using spatial markers (external memory cues).
The difference reveals the advance of perceptual (low-level) versus conceptual
(high-level) cognitive processes in a child’s information search, influenced by the
emotional state and cognitive abilities. The study emphasizes the interactive roles
of motivation, spatial cognition, and information-seeking strategies in digital
contexts (Wu & Chen, 2016). In the other studies the integration of curiosity-
driven mechanisms into machine learning models to enhance their efficiency and
adaptability, it introduces frameworks such as the Contrastive-Curiosity-Driven
Learning Framework (CCLF), Curiosity-Driven Unsupervised Data Collection
(CUDC), and Curiosity-Driven Single Hidden Layer Feedforward Network (CD-
SLFN), aiming to improve reinforcement learning and classification tasks by
incorporating intrinsic motivation inspired by human curiosity.
Positive and Negative Impact:
As we are approaching the globalization era, the importance to integrate and
retrieve the information is more needed than ever. Particularly, the need for timely
and accurate information has led to the application of different strategies to
improve information-seeking behaviors. Generative AI Applications (Nahla &
Masruri, 2024) One of the most extant strategies. Chatbots and natural language
processing models are some examples of generative AI tools that process large
datasets and present quick and easy-to-understand results. This has changed the
way university students, professionals, and general users searching for
information.
In this paper, Nahla and Masruri (2024) give a detailed account of the contribution
of generative AI to information-seeking behavior, explaining the positive and
negative consequences of AI application. On the positive side, they highlight that
generative AI dramatically accelerates the information retrieval process. For
instance, AI tools can minimize the time that it would take to navigate through
huge data sets, returning concise replies to user inquiries. In the academic world
where students and researchers are sometimes required to find information in a
time-efficient manner, this is an added advantage. Other studies, such as those by
Chaudhary et al., have matched the efficiency of AI in delivering relevant
information. (2024), who stress that AI-driven platforms enhance the speed and
accuracy of academic research.
But Nahla and Masruri (2024) also note the potential drawbacks of overreliance on
generative AI for information-seeking. One of the main issues is the accuracy of
AI-generated data. Although AI tools utilize sophisticated algorithms, they are far
from perfect and can occasionally provide incorrect or misleading information
(Qiu et al.). (2024). This fact can be very problematic, especially when it comes to
academic research, where the integrity of information is important. The study notes
yet another downside, dependency on technology.
Dependence on AI to sate users' information needs might impede users' critical
thinking and problem-solving skills, instead of engaging more with the subject
matter (The Guardian, 2025), given the fact that AI may also allow users to skip
most of the deeper engagement with it. Singh and Manjaly (2021) note that the
high availability of AI-generated material will result in receptivity to superficial
learning instead of deep critical engagement. Nahla and Masruri (2024)
acknowledge these challenges and urge regulations and oversight to address the
risks of using AI for information retrieval. They emphasize a balanced approach
that integrates AI tools to complement the educational process while users still
recognize the constraints and risks it poses. This aligns with Hou et al. (2024) who
suggest integrating AI tools into (learning) environments in a way that provides
appropriate guidance and support for users to acquire the skills and competences
needed to evaluate and use the information critically.
Overall, generative AI is undoubtedly a boon for information retrieval, making it
more rapid and user-friendly, however, its drawbacks, particularly concerning data
accuracy and overdependence on tech, shouldn't be dismissed. We should promote
healthy AI use that preserves and emphasizes critical thinking requirements
without hindering the academic and, therefore, intellectual development of our
students, while keeping a close eye on advancing AI development up through
October of 2023 or any remotely similar construct.
Wang and Li (2024) expanded the Expectation-Confirmation Model by
incorporating cognitive and affective variables and explored their impact on
students' autonomous learning when utilizing AI. The study determined that
positive emotions and digital efficacy are positively correlated to student
satisfaction and continued and frequent use of AI tools, and digital efficacy
impacts the extent to which students are willing to learn autonomously. Most
importantly, it highlights the need for positive emotional experiences and digital
competency to support students being self-directed learners.
Research Gap:
A primary gap is exploration of AI’s long-term impact on curiosity. Existing
studies have examined the more immediate positive and negative impacts of AI-
enabled personalized information seeking (Bawden & Robinson, 2009; Pariser,
2011) but they do not inform the long-term implications of AI for curiosity. While
most studies examine the short-term effects of AI (e.g. user engagement or task
completion), the long-term implications of persistent reliance on AI for fostering
users’ internal motivation to learn new ideas and knowledge over time yet to be
addressed. More research is needed to examine how AI technologies shape
curiosity over time, especially in educational contexts where lifelong learning is
crucial. Another gap consists in a lack of comparative studies of curiosity between
domains. Curiosity is a well-known driver of learning (Silvia, 2008), but the
extent to which curiosity motivates information-seeking in different domains (e.g.,
academic vs. personal learning or professional vs. leisure exploration) has not been
systematically explored.
Curiosity as it occurs in academic environments, where the acquisition of facts is
often motivated by specific learning goals, may differ dramatically from the way
curiosity motivates exploration in personal contexts, where the acquisition of facts
is more often motivated by intrinsic interest. Grasping these domain-specific
differences can help drive the design of AI systems which match diverse curiosity
driven behaviors across different domains.
There’s also an interaction between AI and cognitive load that is underexplored.
Tools of AI aimed at helping people do information seeking like Lazarsfeld
Google recommendations and engines help reduce cognitive load by simply
providing information in a relatively personalized way (Russell Norvig, 2016). Yet
no research exists on the effects this lessening of cognitive noise has on curiosity
and deep learning.
Would AI’s streamlined process for information-seeking become a deterrent to
cognitive ‘deepening’ skills, like critical thinking, or promote a passive learning
environment that could lessen curiosity as we go along? Future studies may
investigate the effects of AI on not only how much information users' access, but
also how deeply users process and explore information.
Moreover, we have a blind spot when it comes to the role of AI in our development
of a variety of curiosity. Recommendation algorithms tailored to individual users
often exacerbate pre-existing preferences, restricting users to a narrow scope of
information (Pariser, 2011). But there has been less attention on how AI systems
can be designed to elicit a wider range of curiosity. In this regard, a second area of
research can investigate ways in which AI tools can be engineered to expose,
rather than reinforce, user preferences so that users are exposed to novel,
unexpected, or challenging content that widens their horizons. This is especially
salient in educational technologies where encouraging diverse models of curiosity
could drive unique learning pathways.
A second gap relates to the impact of AI on uncertainty and information seeking.
Although curiosity is often motivated by a sense of uncertainty (Loewenstein,
1994), little research has examined how AI systems diminish or amplify the
feeling of uncertainty creating a desire to know more. While AI can offer simple
answers and reduce vagueness, it can also oversimplify or give information out of
context, leading to a lack of intrigue that comes from the natural mystique of
uncertainty. Future work may study the balance between providing information by
AI systems but also maintaining a level of uncertainty that fosters curiosity-driven
information exploration.
Considerable research has emerged examining how AI impacts user behavior,
however, little to none has been released about the implications of ethical AI and
how it is shaping curiosity and information (or lack thereof) seeking. For instance,
how do AI algorithmic biases affect what information gets shown to users? Might
AI amplify existing social biases or limit exposure to a variety of perspectives,
thereby shaping how users think and fueling — or stifling — their curiosity? While
algorithmic filters can lead to manipulation (e.g., by encouraging specific political
ideologies), the ethics of using this knowledge is particularly concerning in an
educational context where the idea is to encourage open-mindedness.
While most of the research to date has centered around AI’s impact on curiosity
and information seeking, less is known about the extent to which an individual’s
level of curiosity is associated with the adoption and efficacy of AI tools.
Individuals who score high on curiosity might be more inclined to experiment with
the breadth of AI-powered systems while those who are less curious stick with
basic functions the research could set the stage for designing AI systems that
incentivize more active interaction and wider user participation.
Furthermore, a great deal of the AI and information seeking research has been
conducted on Western, tech-savvy populations, with a lack of further investigating
cross-cultural differences. The use of AI tools may also be influenced globally by
cultural differences in information-seeking behavior and curiosity. Some cultures
are more curious than others, which might shape their implementation of AI
solutions; some cultures have limited access or adoption to AI tools; research
could explore how to adapt AI systems specifically for those populations. This
research will provide insights to design culturally relevant AI technologies, which
can better serve global learners.
Finally, there is a lack of research into how AI influences curiosity in the older
population. While many articles focus on younger or digital-first user (Gruber et
al., 2014; Heffernan & Heffernan, 2014), there is little research talking about AI's
effect on curiosity in older populations. As AI solutions are being integrated into
the online ecosystem, determining how the elderly population uses such tools to
gather information and explore their curiosity could suggest how to design and
implement solutions that will be more accessible and engaging for the group.
Studies could also explore the relationship between AI driven information seeking
and cognitive aging.
Theoretical Framework:
1.Self-Determination Theory (SDT) (Deci & Ryan, 1985):
SDT suggests that human motivation is driven by the need to fulfill three basic
psychological needs: autonomy, competence, and relatedness.
In the context of AI, the use of AI tools can potentially support or hinder these
needs. If AI tools help students gain knowledge efficiently, they may feel more
competent, thus increasing curiosity. However, excessive reliance on AI could
potentially undermine autonomy, as students might become overly dependent on
AI-generated information, reducing intrinsic motivation to explore and learn
independently.

2.The Information Foraging Theory (Pirolli & Card, 1999):

Information Foraging Theory compares information-seeking behavior to animal
foraging. It suggests that individuals act like information "hunters," optimizing the
cost-benefit ratio of acquiring information.
AI tools can enhance information-seeking behavior by providing better access to
relevant information, reducing the "cost" of information retrieval, and increasing
the "reward" of acquiring valuable knowledge. This could lead to more effective
and efficient curiosity-driven behavior.

3.Cognitive Load Theory (Sweller, 1988):

Cognitive Load Theory posits that learning is affected by the load placed on
working memory. Effective learning occurs when cognitive load is optimized.
AI can assist in reducing cognitive load by providing concise, tailored information
that helps students focus on higher-order thinking and exploration. However, AI
might also increase cognitive load if students need to navigate complex interfaces
or process an overwhelming amount of data.

4.The Flow Theory (Csikszentmihalyi, 1990):

Flow occurs when a person is deeply engaged in an activity that is both
challenging and rewarding, with clear goals and immediate feedback.
AI tools that provide immediate feedback and adapt to the user’s skill level can
facilitate a state of flow, potentially fostering increased curiosity and sustained
information-seeking behavior. If AI is tailored to students’ needs and challenges, it
may help maintain an optimal level of engagement.

5.Technological Acceptance Model (TAM) (Davis, 1989):

TAM suggests that perceived ease of use and perceived usefulness determine users'
acceptance of technology.
Students’ willingness to use AI technologies for information seeking is likely
influenced by how easy and useful they perceive the technology to be. This can
affect their curiosity and willingness to explore new information through AI.

6.Curiosity Theory (Loewenstein, 1994):

Curiosity is a motivational state triggered by information gaps. The desire to
reduce uncertainty drives individuals to seek out new knowledge and experiences.
AI can be seen as both a tool to satisfy curiosity and as a potential source of
curiosity itself, as students may be drawn to explore AI-driven recommendations
and suggestions. If AI systems generate information gaps or provide new,
unexpected knowledge, they might further stimulate curiosity.

7.Constructivist Learning Theory (Bruner, Piaget, Vygotsky):

Learners actively construct knowledge through experience, exploration, and
reflection
. AI tools like chatbots or adaptive learning platforms can support constructivist
learning by allowing students to explore topics at their own pace and receive
tailored responses, potentially enhancing curiosity and discovery.

8.Metacognition Theory:
Refers to one's awareness and control of their own thinking and learning processes.
AI tools that support reflective learning or give feedback (like ChatGPT or AI
tutors) can improve students’ metacognitive awareness—helping them recognize
what they know and what they need to know, which fuels curiosity and targeted
information-seeking

9.Social Cognitive Theory (Bandura, 1986):

Learning occurs through observation, imitation, and modeling, and is influenced
by self-efficacy.
Students may model their use of AI tools based on peers or influencers. AI can
also affect students' sense of self-efficacy in learning—if it boosts their confidence
in understanding, curiosity may increase.

10.Epistemic Curiosity Theory (Berlyne, 1954)

Differentiates between diversive curiosity (desire for novelty) and specific
curiosity (desire to solve a particular problem).

AI tools may trigger either form of curiosity. For instance, exploratory tools like
AI-based content recommendations could spark diversive curiosity, while
question-answering features satisfy specific curiosity.

11.Uses and Gratifications Theory (Blumler & Katz, 1974):

People actively choose media/tools based on the gratifications they seek (e.g.,
information, entertainment, social interaction).
This can be applied to understand why students use AI tools (e.g., for quick
information vs. deep learning), and how this affects their curiosity and
engagement.

12.Cognitive Evaluation Theory (a sub-theory of SDT):

 External events (e.g., rewards, feedback) influence intrinsic motivation by
affecting perceptions of competence and autonomy.
How AI gives feedback (e.g., instant answers vs. guided discovery) can enhance or
diminish curiosity, depending on whether it supports autonomy and perceived
mastery.

13.Cognitive Flexibility Theory (Spiro et al., 1987):

Emphasizes the need for learners to adapt their thinking to new, complex, and ill-
structured domains.
AI tools that present information in multiple formats (text, video, interactive) can
support cognitive flexibility and stimulate curiosity in complex subjects.