Are Teachers Addicted to AI?

Analyzing Factors Influencing Dependence on Generative AI

Through the I-PACE Model

Yiran Du1*
1
University College London, Institute of Cognitive Neuroscience, London, UK
Email: dyr20000801@[Link]
ORCiD: [Link]

*Corresponding author
Yiran Du
1402 Hill House
17 Highgate Hill
London, UK, N19 5FG
Email: dyr20000801@[Link]

Abstract
The integration of generative artificial intelligence (AI) into education has revolutionized teaching practices,
offering educators advanced tools for lesson planning, content creation, personalized learning, and
administrative automation. While AI enhances efficiency and instructional effectiveness, concerns have
emerged regarding teachers’ potential over-reliance on these technologies, leading to AI addiction. This
study applies the I-PACE model (Interaction of Person-Affect-Cognition-Execution) to explore the
psychological and behavioral mechanisms underlying teachers’ dependence on generative AI. Using survey
data from 1750 teachers, the study examines factors such as self-efficacy, need for cognition, mood
regulation, positive affect, perceived usefulness, and cognitive absorption in shaping AI addiction. Findings
indicate that cognitive absorption is the strongest predictor of AI dependence, while perceived usefulness,
self-efficacy, and positive affect contribute indirectly through reinforcement mechanisms. Notably, mood
regulation and need for cognition do not significantly influence AI addiction, suggesting that AI
engagement in education is driven more by functional efficiency than emotional dependence. The results
highlight the importance of fostering mindful AI integration in teaching to prevent habitual over-reliance.
This study provides theoretical contributions by extending the I-PACE model to the context of AI addiction
in education and offers practical insights for educators, institutions, and policymakers in promoting
responsible AI use while maintaining teachers' professional autonomy and cognitive engagement.

Keywords
Generative AI, AI addiction, I-PACE model, educational technology, teacher

1. Introduction
The integration of generative artificial intelligence (AI) into education has significantly transformed
teaching practices, offering educators powerful tools for lesson planning, content creation, personalized
learning, and assessment automation (Baı̇ Doo-Anu & Owusu Ansah, 2023). These AI-driven tools allow
teachers to streamline administrative tasks, generate instructional materials, and enhance student
engagement through adaptive learning technologies (Chan & Tsi, 2024). As AI continues to evolve, its role
in education is expected to expand, making it an indispensable tool for modern educators (Chiu, 2024).

However, alongside these benefits, concerns have emerged regarding the potential over-reliance and
addictive tendencies associated with AI use among teachers (Bower et al., 2024). The efficiency,
convenience, and problem-solving capabilities of generative AI can lead to habitual engagement, where
teachers increasingly depend on AI for routine decision-making and instructional support (Zhai, 2024). This
phenomenon may contribute to cognitive offloading, where educators reduce their direct engagement in
creative lesson planning and problem-solving, instead relying on AI to generate content and insights (Celik
et al., 2022). Over time, this habitual reliance may transition into AI addiction, characterized by difficulty
disengaging from AI tools, excessive dependence on AI-generated outputs, and diminished autonomy in
instructional decision-making (Acosta-Enriquez et al., 2025). While previous research has extensively
examined digital addiction in contexts such as social media, gaming, and smartphone use, little attention
has been given to AI-specific addiction, particularly in professional environments like education (Meng et
al., 2022). Unlike passive digital consumption, AI engagement involves interactive, problem-solving, and
knowledge-generation processes, making its addictive potential unique (Suva & Bhatia, 2024; Zhong et al.,
2024). Teachers experiencing deep cognitive absorption and emotional reinforcement from AI use may
struggle to regulate their engagement, reinforcing habitual dependency patterns. Given the growing
adoption of AI-powered tools in education, there is a critical need to examine how dependent teachers have
become on generative AI and what factors contribute to this dependence.

To address these concerns, this study applies the I-PACE model (Interaction of Person-Affect-Cognition-
Execution) to explore the mechanisms underlying generative AI addiction in educators. The I-PACE
framework suggests that technology addiction arises from the interplay of personal traits, cognitive and
affective responses, and reinforcement mechanisms, which sustain habitual use over time (Brand et al.,
2016, 2019). By analyzing key psychological and behavioral factors, such as self-efficacy, need for
cognition, mood regulation, positive affect, perceived usefulness, and cognitive absorption, this study aims
to uncover the pathways leading to AI dependence among teachers. To achieve this, the study seeks to
answer the following research questions:

RQ1: What factors contribute to teachers' addiction to generative AI?

RQ2: To want extent can I-PACE model explain teachers' addiction to generative AI?

The findings of this research will provide valuable insights into the mechanisms and drivers of AI
dependence in education, offering guidance on how teachers can use AI responsibly without becoming
overly reliant on it. Additionally, the study will inform educational institutions and policymakers on
strategies to promote balanced AI integration and digital well-being among educators. By investigating AI
addiction through the I-PACE model, this research contributes to a deeper understanding of how generative
AI impacts teachers' work habits, decision-making, and long-term engagement with AI-driven tools.

2. Literature Review
2.1. Teacher Behavior in Generative AI Use
The rapid advancement of generative AI has transformed multiple industries, including education, where
AI-powered tools assist in content creation, lesson planning, assessment, and personalized learning (Wen
et al., 2025). Teachers increasingly integrate AI to enhance instructional efficiency, support student
engagement, and streamline administrative tasks (Kohnke & Moorhouse, 2025). With AI models capable
of generating customized teaching materials, answering subject-specific queries, and providing real-time
feedback, educators benefit from enhanced productivity and improved access to pedagogical resources (L.
Hu et al., 2025). Despite these advantages, prolonged engagement with generative AI may lead to
unintended behavioral consequences. Teachers who frequently rely on AI may develop habitual usage
patterns, reducing their engagement in independent content creation and decision-making (Baı̇ Doo-Anu &
Owusu Ansah, 2023). The ease of AI-driven assistance may also contribute to cognitive offloading, where
educators delegate tasks to AI rather than actively engaging in problem-solving and instructional design
(Kohnke & Moorhouse, 2025). Over time, excessive dependence on AI tools may lead to a shift from
functional use to habitual reliance, raising concerns about potential AI addiction in educational settings
(Bozoglan et al., 2014).

Understanding teacher behavior in AI use requires examining both the motivational factors that drive AI
adoption and the psychological mechanisms that sustain continued engagement. Research suggests that
teachers’ self-efficacy, cognitive tendencies, and emotional responses influence how they integrate AI into
their work (Bower et al., 2024; Celik et al., 2022; L. Hu et al., 2025). However, while prior studies have
explored AI adoption and perceived benefits, limited research has investigated the risks of over-reliance
and compulsive AI use among educators (Chan & Tsi, 2024; Wen et al., 2025). As generative AI continues
to evolve, there is an urgent need to analyze the underlying mechanisms that contribute to AI dependency
and ensure that AI remains a supportive tool rather than a disruptive force in education.

2.2. Generative AI Addiction

AI addiction, a form of technology overuse characterized by compulsive engagement and reduced self-
regulation, has gained attention as AI systems become more sophisticated and interactive (Y. Huang &
Huang, 2024). Unlike traditional software applications, generative AI possesses adaptive learning
capabilities, personalized interaction features, and real-time problem-solving abilities, making it
particularly immersive and engaging (Bower et al., 2024; Brand et al., 2019). As AI tools become more
accessible, concerns have emerged regarding excessive AI dependence and its impact on users’ cognitive,
emotional, and behavioral patterns.

Generative AI addiction shares similarities with internet and social media addiction, where users develop
strong emotional attachments, deep cognitive absorption, and habitual overuse. The defining features of AI
addiction include prolonged engagement, difficulty in disengaging, reliance on AI for daily tasks, and
neglect of real-world problem-solving skills. In an educational context, teachers who become excessively
dependent on AI may struggle to complete tasks without AI assistance, experience reduced creativity, and
exhibit diminished engagement in traditional instructional planning (Comiskey, 2024). Several
psychological factors contribute to AI addiction, including cognitive immersion, emotional reinforcement,
and perceived usefulness (Brand et al., 2016, 2019). The convenience of AI-generated content, combined
with instant feedback and adaptive learning capabilities, reinforces continued engagement, creating habit-
forming behavioral loops (Renshaw & Carley, 2024). Over time, teachers who develop strong cognitive
and affective connections with AI may find it challenging to regulate their usage patterns, leading to over-
reliance and potential addiction (Dinnie, 2024).

Existing studies on digital addiction have primarily focused on gaming, social networking, and smartphone
use, with limited research on AI-specific addiction behaviors (Galvan & Newman, 2025; Li et al., 2025;
Piko et al., 2025). As generative AI becomes increasingly integrated into educational systems,
understanding the risk factors and behavioral patterns associated with AI dependency is essential for
developing strategies to promote responsible AI use (Almulla et al., 2025). This study seeks to address this
gap by exploring the psychological mechanisms underlying AI addiction among teachers, providing
insights into how educators can harness AI effectively while mitigating the risks of excessive reliance.

2.3. I-PACE Model

The I-PACE model (Interaction of Person-Affect-Cognition-Execution) provides a comprehensive
framework for understanding the development of addictive behaviors in digital environments (Brand et al.,
2016, 2019). It proposes that compulsive technology use emerges from the interplay between personal
predispositions, cognitive and affective responses, and reinforcement processes over time (Lörsch et al.,
2025). Originally designed to explain internet gaming disorder, the model has since been widely applied to
various forms of digital addiction, including social media overuse, smartphone addiction, and AI-driven
interactions (Calaresi et al., 2024; Tie et al., 2025). By focusing on how psychological mechanisms shape
technology engagement, the I-PACE model offers a dynamic perspective on behavioral addiction,
distinguishing between initial engagement, habitual use, and compulsive dependency (Pupi et al., 2024;
Zhu & Fong, 2025).

In the context of generative AI adoption in education, the I-PACE model serves as an essential theoretical
lens for examining teacher dependency on AI tools. Generative AI, with its ability to generate text, automate
lesson planning, provide personalized content, and enhance instructional efficiency, has become a valuable
asset in modern education (Y. Huang & Huang, 2024). However, as AI tools become more sophisticated,
concerns arise regarding over-reliance and excessive use (Suva & Bhatia, 2024). Teachers who frequently
turn to AI for assistance may experience a gradual shift from functional use to habitual engagement,
eventually forming patterns of dependency and addiction (Comiskey, 2024).

The I-PACE framework explains this process by identifying three primary components that contribute to
technology addiction (S. Zhang et al., 2024). Personal characteristics, such as self-efficacy and need for
cognition, influence how individuals initially approach technology, shaping their motivation, engagement,
and confidence in using AI tools (Brand et al., 2019). Affective and cognitive responses, such as mood
regulation, positive affect, perceived usefulness, and cognitive absorption, further reinforce usage patterns
(Brand et al., 2019). For example, teachers who use AI to reduce stress and enhance productivity may
experience positive emotional reinforcement, making them more likely to continue and intensify AI
engagement (Bower et al., 2024). Over time, these cognitive and emotional responses can create
reinforcement loops, where the ease and efficiency of AI-driven interactions decrease self-regulation and
increase dependency. This leads to sustained engagement, immersion, and, in some cases, compulsive AI
use, disrupting normal teaching practices and professional decision-making. Unlike traditional models that
focus only on technology adoption or frequency of use (Cheung & Vogel, 2013; Dehghani & Mashhadi,
2024; Hsu & Lin, 2022), the I-PACE model highlights the underlying psychological mechanisms that drive
sustained engagement and eventual dependency (S. Zhang et al., 2024).

Applying the I-PACE framework, this study seeks to uncover the factors contributing to generative AI
addiction among teachers. By examining how personal, cognitive, and affective factors interact to shape AI
dependency, this research aims to provide insights into how educators can integrate AI effectively without
developing over-reliance. The findings will also inform educational institutions on how to promote balanced
and mindful AI adoption, ensuring that AI serves as an enhancing tool rather than a replacement for critical
thinking and decision-making.

3. Hypothesis development
The I-PACE model (Interaction of Person-Affect-Cognition-Execution) provides a structured approach to
understanding how individual characteristics, cognitive and emotional responses, and reinforcement
mechanisms contribute to addictive behaviors in digital environments. This study applies the I-PACE
framework to investigate how personal factors (self-efficacy and need for cognition), affective factors
(mood regulation and positive affect), and cognitive factors (perceived usefulness and cognitive absorption)
interact to shape generative AI addiction among teachers. The proposed research model is illustrated in
Figure 1.

3.1. Self-Efficacy
Self-efficacy refers to an individual’s belief in their ability to successfully perform tasks using a particular
system (Trusz & Babel, 2016). In the context of generative AI, teachers with high self-efficacy are likely to
be confident in their ability to integrate AI into their professional workflows (Fryer et al., 2020). Previous
research has shown that self-efficacy enhances technology adoption, learning outcomes, and engagement,
as individuals with greater confidence in their skills are more willing to explore and experiment with AI-
powered tools (Bewersdorff et al., 2025; Rodríguez-Ruiz et al., 2025).

A strong sense of self-efficacy can lead to greater cognitive immersion, as teachers who feel capable of
using AI effectively are more likely to become absorbed in interactions with the technology (Liang et al.,
2023). Furthermore, self-efficacy may enhance the perceived usefulness of generative AI, as confident users
are more likely to view AI as a valuable tool that complements their expertise (Yao & Wang, 2024).
Additionally, self-efficacy may contribute to mood regulation, as teachers who trust their ability to use AI
effectively may experience lower frustration and greater emotional stability during AI-assisted tasks (Y.
Wang et al., 2022). Thus, we propose the following hypotheses:
H1a. Self-efficacy positively influences mood regulation.
H1b. Self-efficacy positively influences cognitive absorption.
H1c. Self-efficacy positively influences perceived usefulness.

3.2. Need for Cognition

Need for cognition is a personality trait that reflects an individual's tendency to seek out, engage in, and
enjoy complex cognitive tasks (Cacioppo & Petty, 1982). Teachers with a high need for cognition are more
likely to explore AI-generated content, engage in deep analytical thinking, and appreciate AI’s ability to
process large amounts of information efficiently. Given that generative AI provides opportunities for
exploration, problem-solving, and critical thinking, teachers with a strong need for cognition may find
themselves increasingly immersed in AI-driven tasks. (Tanas et al., 2020) This cognitive inclination may
lead to higher cognitive absorption, as teachers with intellectual curiosity may spend extended periods
interacting with AI to satisfy their desire for knowledge and problem-solving. Over time, this heightened
engagement may also contribute to generative AI addiction, as deep immersion and habitual reliance on AI
become reinforcing mechanisms for continued use. Teachers who frequently turn to AI for intellectual
stimulation may develop patterns of excessive engagement, leading to compulsive AI use and potential
difficulty in disengaging from AI-assisted tasks. Based on these insights, we propose the following
hypotheses:

H2a. Need for cognition positively influences generative AI addiction.

H2b. Need for cognition positively influences cognitive absorption.

3.3. Mood Regulation

Mood regulation refers to the use of external stimuli to manage emotional states, helping individuals cope
with stress, anxiety, and cognitive overload (Larsen, 2000). Teachers often face high workloads, time
constraints, and complex decision-making tasks, making AI an appealing tool for emotional relief and stress
reduction (Qu & Wang, 2024). By automating repetitive tasks such as lesson planning, grading, and content
generation, generative AI can provide a sense of efficiency and control, alleviating teachers' emotional
burdens and reinforcing habitual AI engagement (Dehghani & Mashhadi, 2024). As teachers increasingly
rely on AI for mood regulation, they may develop stronger perceptions of its usefulness, viewing it as an
essential tool for reducing workload, enhancing efficiency, and improving decision-making. Furthermore,
frequent AI use for emotional relief may lead to cognitive absorption, where teachers become deeply
immersed in AI interactions and spend extended periods engaging with AI-powered tools. Over time, this
dependence on AI for stress relief and efficiency may also contribute to generative AI addiction, as teachers
find it difficult to disengage from AI-driven workflows Based on these insights, we propose the following
hypotheses:

H3a. Mood regulation positively influences perceived usefulness.

H3b. Mood regulation positively influences cognitive absorption.
H3c. Mood regulation positively influences generative AI addiction.

3.4. Positive Affect

Positive affect reflects pleasurable emotional responses that arise from engaging in an activity, such as joy,
excitement, and satisfaction (Dockray & Steptoe, 2010). When teachers experience positive emotions while
using AI, they are more likely to develop a stronger attachment to the tool and engage with it more
frequently (S. Wang et al., 2023). Enjoyable interactions with AI can foster deeper involvement, increasing
the likelihood of sustained engagement and immersion in AI-assisted tasks (F. Huang et al., 2024). Teachers
who derive positive emotional experiences from AI use may also perceive it as more useful, reinforcing
their belief that AI enhances their productivity, efficiency, and instructional effectiveness. Furthermore,
positive affect may contribute to cognitive absorption, where teachers become so engaged with AI
interactions that they lose track of time and remain deeply focused on AI-generated outputs (Liu & Chang,
2024). Additionally, the emotional benefits of AI use may support mood regulation, as teachers use AI not
only for professional efficiency but also for emotional comfort in managing work-related stress (S. Wang
et al., 2023). Based on these insights, we propose the following hypotheses:

H4a. Positive affect positively influences perceived usefulness.

H4b. Positive affect positively influences cognitive absorption.
H4c. Positive affect positively influences mood regulation.

3.5. Perceived Usefulness

Perceived usefulness reflects the extent to which individuals believe that a technology enhances their
productivity, efficiency, and overall work performance (Zou et al., 2024). In an educational setting, teachers
are more likely to integrate AI into their workflows if they perceive it as beneficial for lesson planning,
research, content creation, and classroom management. When AI streamlines tasks and simplifies complex
instructional processes, teachers may experience greater engagement and reliance on AI-driven support
(Bower et al., 2024). As teachers increasingly recognize the practical benefits of AI, they may also
experience higher levels of cognitive absorption, becoming deeply immersed in AI-assisted tasks (C. Zhang
et al., 2023). The more indispensable AI appears in their professional routines, the more likely teachers are
to spend extended periods engaging with AI tools, reinforcing a state of deep concentration and interaction.
This suggests that perceived usefulness plays a significant role in enhancing cognitive absorption, as
teachers who see AI as highly beneficial may find themselves completely engaged and focused on AI-
powered tasks (Dehghani & Mashhadi, 2024). Based on these insights, we propose the following hypothesis:

H5. Perceived usefulness positively influences cognitive absorption.

3.6. Cognitive Absorption

Cognitive absorption refers to deep mental engagement and immersion in an activity, where individuals
become so engrossed that they lose track of time and external surroundings (Agarwal & Karahanna, 2000).
In the context of generative AI, teachers who experience high cognitive absorption may spend excessive
amounts of time interacting with AI tools, fully immersing themselves in AI-driven tasks (Saadé & Bahli,
2005). As teachers engage deeply with AI, they may develop habitual usage patterns that lead to increasing
reliance on AI-generated outputs for decision-making, content creation, and instructional planning. Over
time, this deep immersion may evolve into dependency, where teachers feel unable to disengage from AI
tools, even when excessive use begins to interfere with their professional autonomy and independent
problem-solving skills (Lin, 2009). This suggests that cognitive absorption plays a crucial role in
reinforcing generative AI addiction, as prolonged and immersive engagement may lead to compulsive AI
use. Based on these insights, we propose the following hypothesis:

H6. Cognitive absorption positively influences generative AI addiction.

Figure 1. The proposed model

4. Method
4.1 Research instrument
This study employs a structured research instrument to examine the factors influencing teachers'
dependence on generative AI, based on the I-PACE model (Interaction of Person-Affect-Cognition-
Execution) (Brand et al., 2016, 2019). The research framework includes seven key constructs, each
representing different psychological and behavioral aspects of AI engagement in education. These
constructs were identified through a review of existing literature, and their corresponding measurement
items were adapted from validated sources to ensure content validity. A five-point Likert scale, ranging
from 1 (strongly disagree) to 5 (strongly agree), was used to measure all items, providing a robust
assessment of variations in participants’ perceptions and behaviors. To refine the questionnaire and enhance
clarity, a pretest was conducted with 10 educators who frequently use generative AI tools in their teaching
practices. These participants reviewed the questionnaire and provided feedback regarding the clarity,
relevance, and comprehensibility of the items. Based on their input, minor modifications were made to
improve the readability of certain statements and ensure that they effectively captured the intended
constructs. This pretesting process helped to enhance the reliability of the measurement instrument before
its full-scale deployment (Du, 2024).

The final research instrument consists of seven constructs, each reflecting a specific dimension of AI
engagement in teaching. Self-Efficacy (SE) assesses teachers' confidence in their ability to effectively use
generative AI for instructional tasks, including troubleshooting and adapting AI-generated insights. Need
for Cognition (NC) measures teachers' intellectual curiosity and motivation to explore AI-driven
innovations for lesson planning and pedagogical improvement. Mood Regulation (MR) evaluates how
teachers use AI as a tool to manage work-related stress, reduce workload pressure, and maintain emotional
well-being. Positive Affect (PA) captures the emotional benefits teachers experience, such as excitement,
motivation, and a sense of accomplishment when integrating AI into their teaching strategies. Furthermore,
Perceived Usefulness (PU) examines teachers' beliefs about the practical benefits of AI, particularly in
improving efficiency, productivity, and instructional quality. Cognitive Absorption (CA) assesses the extent
to which teachers become deeply immersed in AI-based tasks, often losing track of time while working
with AI-generated content. Lastly, Generative AI Addiction (GA) investigates excessive reliance on AI tools,
including difficulties in limiting AI use, potential disruptions to teaching autonomy, and signs of compulsive
engagement. The measurement items for each construct, along with their sources, are presented in Table 1.

Table 1. Constructs and measurement items

Construct Item Item Content Source
Self-Efficacy SE1 I feel confident in my ability to use generative AI tools to (Y.-Y. Wang &
(SE) support my teaching tasks. Chuang, 2024)
SE2 I can independently figure out how to use new features of
generative AI in my educational activities.
SE3 I believe I can troubleshoot issues that arise when using
generative AI in my teaching.
Need for NC1 I enjoy exploring how generative AI can improve my (Lord &
Cognition lesson plans and teaching strategies. Putrevu, 2006)
(NC) NC2 I often use generative AI to find innovative ways to explain
complex concepts to my students.
NC3 I feel intellectually stimulated when I discover new
applications of generative AI in education.
Mood MR1 I use generative AI tools to reduce the stress of preparing (Hutchison &
Regulation lesson materials. Gunthert,
(MR) MR2 When I feel overwhelmed with tasks, I turn to generative 2013)
AI to ease my workload.
MR3 Using generative AI helps me stay calm and focused when
I have a lot to do.
Positive PA1 I feel excited when I successfully create valuable teaching (Joshanloo,
Affect (PA) resources using generative AI. 2017)
PA2 Using generative AI tools makes me feel more motivated to
experiment with new teaching methods.
PA3 I experience a sense of accomplishment when I integrate
AI-generated content into my lessons.
Perceived PU1 Generative AI tools help me prepare high-quality teaching (Zou et al.,
Usefulness materials more efficiently. 2023)
(PU) PU2 Using generative AI allows me to focus more on engaging
with my students rather than on routine tasks.
PU3 Generative AI provides me with valuable ideas and
resources that I wouldn’t have thought of on my own.
PU4 I find generative AI to be a useful tool for improving my
productivity as a teacher.
Cognitive CA1 I often lose track of time when I use generative AI tools to (Bozoglan et
Absorption prepare my lessons. al., 2014)
(CA) CA2 When I use generative AI for teaching tasks, I become so
absorbed that I don’t notice other things happening around
me.
CA3 I find myself fully focused and engaged when working with
generative AI on educational content.
Generative GA1 I often feel that I use generative AI tools more than I should (B. Hu et al.,
AI Addiction for my teaching tasks. 2023)
(GA) GA2 I have tried to limit my use of generative AI tools, but I
find it difficult to do so.
 GA3 My reliance on generative AI sometimes interferes with my
ability to complete tasks without it.
GA4 I feel uneasy or frustrated when I am unable to use
generative AI tools.

4.2 Data collection

The survey was conducted in a city in China from January 15, 2025, to February 10, 2025, targeting teachers
who actively use or have experience with generative AI in their teaching practices. A structured online
questionnaire was distributed through professional teaching networks, educational forums, and institutional
email lists to ensure broad participation. To enhance the reliability and validity of responses, participants
were assured of anonymity, and informed consent was obtained before data collection. Over the period,
1750 valid responses were collected after a rigorous data-cleaning process. Careless responses (Ward &
Meade, 2023), as well as those from participants who had never used or were completely unfamiliar with
generative AI, were removed to ensure data quality. The final dataset reflects a diverse and representative
sample of educators in the city, allowing for a comprehensive examination of the factors influencing
teachers' dependence on generative AI.

4.3 Data analysis

The data analysis involved a structured approach using both descriptive and inferential statistical techniques
to examine the relationships between key constructs based on the I-PACE model. Initially, data cleaning
and preprocessing were conducted to ensure accuracy and reliability. Descriptive statistics provided insights
into demographic distributions and AI usage patterns among teachers. Confirmatory factor analysis (CFA)
was used to assess the validity and reliability of the measurement model, ensuring that the constructs were
well-defined and internally consistent. Structural equation modeling (SEM) was employed to test the
hypothesized relationships among variables, examining the direct and indirect effects of psychological and
behavioral factors on AI dependence. Model fit indices were evaluated to confirm the robustness of the
proposed framework. The analysis aimed to uncover underlying mechanisms driving AI engagement among
educators, offering theoretical and practical implications for managing AI reliance in educational settings.

5. Results
5.1 Descriptive statistics
Table 2 provides an overview of the demographic and behavioral characteristics of the study participants.
The sample consists of 1,750 teachers, with 85.57% identifying as male and 15.43% as female. Age
distribution reveals that 8.29% are below 25 years, 28.51% fall within the 25-34 age group, 29.20% are
aged 35-44, another 29.20% are 45-54, and 4.80% are above 55. In terms of education, 48.74% hold an
associate degree, 49.83% a bachelor’s degree, and 1.43% a master’s degree. The participants’ teaching
experience varies, with 4.46% having less than a year, 13.37% with 1-3 years, 14.00% with 4-6 years, 11.71%
with 7-10 years, and 56.46% with over 10 years. Regarding teaching levels, 9.26% teach kindergarten,
54.86% primary school, 26.69% middle school, and 9.20% high school. AI usage patterns indicate that the
most commonly used AI tool is DeepSeek (64.00%), followed by ERNIE Bot (48.91%), 360 Zhinao
(13.14%), ChatGPT (12.23%), Tencent Hunyuan Assistant (6.69%), iFlytek Spark Cognition Model
(6.51%), Ali Tongyi Qianwen (3.49%), Claude (3.03%), and Huawei Pangu Model (2.23%). AI familiarity
levels show that 85.60% are somewhat familiar, 12.57% are relatively familiar, and 1.83% are very familiar.
The duration of AI use varies, with 30.11% using it for less than a month, 31.37% for 1-3 months, 16.69%
for 4-6 months, 8.74% for 7-12 months, and 13.09% for over a year. The primary purposes for AI use
include lesson preparation (79.49%), content creation such as examination papers (61.71%), student
engagement enhancement (33.31%), marking and evaluation (23.03%), professional development (31.94%),
personal interests or entertainment (34.97%), and other purposes (48.34%). These statistics provide insight
into the diverse backgrounds and AI engagement levels of teachers, highlighting key trends in AI reliance
and its integration into educational practices.
Table 2. Demographic and behavioral characteristics of the study participants
Measure Item Frequency Percent
Gender Female 1480 15.43%
Male 270 85.57%
Age Below 25 145 8.29%
25-34 499 28.51%
35-44 511 29.20%
45-54 511 29.20%
Above 55 84 4.80%
Education Associate degree 853 48.74%
Bachelor’s degree 872 49.83%
Master’s degree 25 1.43%
Teaching years Less than 1 year 78 4.46%
1-3 years 234 13.37%
4-6 years 245 14.00%
7-10 years 205 11.71%
More than 10 years 988 56.46%
Teaching level Kindergarten 162 9.26%
Primary school 960 54.86%
Middle school 467 26.69%
High school 161 9.20%
AI tool used ChatGPT 214 12.23%
Claude 53 3.03%
ERNIE Bot 856 48.91%
Tencent Hunyuan Assistant 117 6.69%
Ali Tongyi Qianwen 61 3.49%
iFlytek Spark Cognition Model 114 6.51%
Huawei Pangu Model 39 2.23%
360 Zhinao 230 13.14%
DeepSeek 1120 64.00%
AI familiarity Somewhat familiar 1498 85.60%
Relatively familiar 220 12.57%
Very familiar 32 1.83%
Duration of use Less than 1 month 527 30.11%
1-3 months 549 31.37%
4-6 months 292 16.69%
7-12 months 153 8.74%
More than a year 229 13.09%
Purpose of use Lesson preparation 1391 79.49%
Creating teaching content (such 1080 61.71%
as examination papers, exercises)
Marking and evaluation 403 23.03%
Increasing student engagement 583 33.31%
Professional development 559 31.94%
Personal interests or 612 34.97%
entertainment
others 846 48.34%

5.2 Measurement model

To ensure the reliability and validity of the measurement model, the study conducts Confirmatory Factor
Analysis (CFA), assessing internal consistency, convergent validity, and discriminant validity. Table 3
presents the reliability and validity results, including factor loadings, Cronbach’s alpha (α), composite
reliability (CR), and average variance extracted (AVE). All factor loadings exceed the recommended
threshold of 0.70, confirming adequate indicator reliability. Cronbach’s alpha values range from 0.86 to
0.96, while composite reliability (CR) values are above 0.86 for all constructs, demonstrating high internal
consistency. Additionally, AVE values range from 0.68 to 0.88, surpassing the minimum threshold of 0.50,
thereby supporting convergent validity. Discriminant validity is assessed using the Fornell-Larcker criterion,
which requires that the square root of the AVE for each construct be higher than its correlations with other
constructs. Table 4 presents the correlation matrix, with the square roots of AVEs displayed along the
diagonal in bold. The results indicate that each construct shares more variance with its own indicators than
with other constructs, confirming discriminant validity. The model fit indices confirm that the measurement
model provides a good fit for the data. As shown in Table 5, the Comparative Fit Index (CFI) is 0.96, and
the Tucker-Lewis Index (TLI) is 0.95, both exceeding the recommended threshold. The Root Mean Square
Error of Approximation (RMSEA) is 0.079, and the Standardized Root Mean Square Residual (SRMR) is
0.05, both meeting acceptable criteria. Overall, the measurement model demonstrates strong psychometric
properties, supporting its reliability, convergent validity, and discriminant validity. These findings provide
a solid foundation for subsequent structural model analysis.

Table 3. Reliability and validity results

Construct Item Loading Cronbach’s α CR AVE
Self-Efficacy (SE) SE1 0.72 0.86 0.86 0.68
SE2 0.88
SE3 0.85
Need for Cognition (NC) NC1 0.88 0.92 0.91 0.79
NC2 0.87
NC3 0.87
Mood Regulation (MR) MR1 0.92 0.95 0.95 0.86
MR2 0.92
MR3 0.94
Positive Affect (PA) PA1 0.92 0.95 0.95 0.87
PA2 0.95
PA3 0.93
Perceived Usefulness (PU) PU1 0.95 0.96 0.97 0.88
PU2 0.90
PU3 0.96
PU4 0.95
Cognitive Absorption (CA) CA1 0.93 0.95 0.95 0.88
CA2 0.94
CA3 0.93
Generative AI Addiction (GA) GA1 0.88 0.93 0.93 0.77
GA2 0.90
GA3 0.88
GA4 0.86

Table 4. Correlation matrix

SE NC MR PU CA PA GA
SE 0.82
NC 0.73 0.88
 SE NC MR PU CA PA GA
MR 0.43 0.78 0.93
PU 0.31 0.70 0.80 0.94
CA -0.13 0.06 0.09 0.33 0.94
PA 0.30 0.68 0.78 0.94 0.15 0.93
GA -0.65 -0.86 -0.84 -0.88 -0.06 -0.92 0.87
Note: The square roots of the AVEs are highlighted in bold along the diagonal

Table 5. Model fit indices

Fix index CFI TLI RMSEA SRMR
Recommended Value > 0.90 > 0.90 < 0.08 < 0.08
Actual Value 0.96 0.95 0.079 0.05

5.3 Structural model

This study builds a structural model to analyze the relationships between the study constructs. The
hypothesis testing results appear in Table 6 and Figure 2. The findings show that Self-Efficacy (SE)
positively influences Mood Regulation (MR) (β = 0.49, p < 0.001), Cognitive Absorption (CA) (β = 0.47,
p < 0.001), and Perceived Usefulness (PU) (β = 0.08, p < 0.01), supporting H1a, H1b, and H1c. However,
Need for Cognition (NC) does not significantly predict Generative AI Addiction (GA) (β = 0.01, p = 0.83)
or Cognitive Absorption (CA) (β = -0.01, p = 0.55), leading to the rejection of H2a and H2b. For Mood
Regulation (MR), it positively influences Perceived Usefulness (PU) (β = 0.20, p < 0.001), supporting H3a,
but does not significantly affect Cognitive Absorption (CA) (β = 0.09, p = 0.09) or Generative AI Addiction
(GA) (β = -0.03, p = 0.65), rejecting H3b and H3c. Positive Affect (PA) positively affects Perceived
Usefulness (PU) (β = 0.74, p < 0.001) and Mood Regulation (MR) (β = 0.64, p < 0.001), confirming H4a
and H4c, but negatively affects Cognitive Absorption (CA) (β = -0.18, p = 0.05), leading to the rejection of
H4b. The results also indicate that Perceived Usefulness (PU) positively influences Cognitive Absorption
(CA) (β = 0.79, p < 0.001), supporting H5, and that Cognitive Absorption (CA) positively affects Generative
AI Addiction (GA) (β = 0.75, p < 0.001), confirming H6. These findings suggest that cognitive absorption
plays a key role in AI dependence, as teachers who become highly engaged with AI tools are more likely
to develop excessive reliance.

Overall, the structural model confirms most of the proposed hypotheses and highlights the psychological
and behavioral mechanisms that lead to AI addiction among teachers. The significant relationships between
self-efficacy, positive affect, perceived usefulness, and cognitive absorption indicate that these factors
contribute to teachers’ continued AI engagement. However, the findings also suggest that need for cognition
and mood regulation do not directly affect AI addiction, implying that other factors may moderate these
relationships. These insights contribute to a deeper understanding of AI dependence in education and
provide guidance on how teachers can balance AI use without excessive reliance.

Table 6. Hypotheses test results

Hypothesis Path Path Coefficient p Result
H1a SE → MR 0.49 *** Supported
H1b SE → CA 0.47 *** Supported
H1c SE → PU 0.08 ** Supported
H2a NC → GA 0.01 0.83 Not Supported
H2b NC → CA -0.01 0.55 Not Supported
H3a MR → PU 0.20 *** Supported
H3b MR → CA 0.09 0.09 Not Supported
H3c MR → GA -0.03 0.65 Not Supported
 H4a PA → PU 0.74 *** Supported
H4b PA → CA -0.18 0.05 Not Supported
H4c PA → MR 0.64 *** Supported
H5 PU → CA 0.79 *** Supported
H6 CA → GA 0.75 *** Supported
Note: Statistical significance is denoted as *** for p < 0.001, ** for p < 0.01, and ** for p < 0.05

Figure 2. Results of structural model testing (**p < 0.01; ***p < 0.001; n.s.: not significant)

6. Discussion
6.1 Factors impacting generative AI addiction
The study’s findings provide key insights into the mechanisms driving generative AI addiction among
teachers. The results confirm that cognitive absorption is the only direct predictor of AI addiction, while
factors such as perceived usefulness, self-efficacy, mood regulation, and positive affect influence AI
engagement indirectly through cognitive absorption. These findings align with and challenge previous
research on technology addiction, offering a clearer understanding of how AI impacts educators' behaviors.

A critical factor in AI addiction is cognitive absorption, which emerged as the strongest predictor of
compulsive AI use. Teachers who become deeply immersed in AI-assisted tasks may find it difficult to
disengage, leading to habitual reliance on AI tools for instructional planning, lesson creation, and
administrative work. This supports previous research on digital immersion and problematic technology use,
which suggests that individuals with high cognitive absorption are more prone to technology overuse
(Agarwal & Karahanna, 2000; Bozoglan et al., 2014). The study also confirms that perceived usefulness
strongly influences cognitive absorption, reinforcing the idea that when educators see AI as an essential
tool, they are more likely to engage deeply with it. This aligns with findings on technology reinforcement
loops, where efficiency gains from AI drive further engagement, making it harder to break the cycle of
dependence (Zhong et al., 2024).

Self-efficacy plays a crucial role in AI engagement, primarily by influencing how teachers interact with AI
rather than directly leading to addiction. Teachers with higher confidence in their ability to use AI tools are
more likely to integrate them into their workflow, experience emotional benefits, and become absorbed in
AI interactions. This finding is consistent with prior studies showing that higher self-efficacy leads to
greater adoption and perceived ease of use of educational technology (Fryer et al., 2020; Bewersdorff et al.,
2025). However, its relatively weaker effect on perceived usefulness suggests that confidence in AI use
does not automatically make teachers perceive AI as indispensable. Instead, its impact is indirect, shaping
teachers' level of engagement rather than compulsive dependence.

Interestingly, need for cognition did not significantly predict AI addiction or cognitive absorption, which
contradicts prior studies that suggest individuals with higher intellectual curiosity engage more deeply with
AI (Lin, 2009; Tanas et al., 2020). This suggests that while teachers with a strong need for cognition explore
AI tools, they do so in a controlled and intentional manner, rather than developing habitual use patterns.
Intellectual curiosity alone does not drive technology addiction; it must be coupled with strong
reinforcement mechanisms such as perceived usefulness or emotional dependence to contribute to
compulsive AI engagement.

Mood regulation, often considered a driver of digital addiction due to its role in stress relief and coping
behaviors, did not show a direct effect on AI addiction. While teachers use AI to manage workload stress
and reduce cognitive overload, this does not necessarily lead to excessive engagement. These results
challenge previous findings that emotional dependence on technology can reinforce compulsive use
patterns (Hu et al., 2023; Ye et al., 2024). Instead, teachers appear to use AI functionally rather than
habitually, treating it as a tool for efficiency rather than as an emotional escape. However, mood regulation
does enhance perceived usefulness, suggesting that teachers who rely on AI to alleviate stress see it as a
more valuable resource in their professional tasks.

The findings on positive affect indicate that while enjoyment from AI use enhances mood regulation and
perceived usefulness, it does not contribute to deeper cognitive absorption. This contradicts the assumption
that pleasurable engagement leads to greater immersion in technology, as seen in previous studies on game
and social media addiction (Dockray & Steptoe, 2010). Instead, it suggests that teachers who experience
enjoyment while using AI are more likely to maintain a balanced approach, engaging with AI strategically
rather than compulsively. This highlights the importance of intrinsic motivation in shaping AI engagement,
where positive affect encourages flexible use rather than overuse.

Taken together, these findings underscore that cognitive absorption is the only direct driver of AI addiction,
while perceived usefulness, self-efficacy, mood regulation, and positive affect indirectly influence addiction
by shaping engagement levels. While self-efficacy facilitates AI adoption, it does not necessarily lead to
addiction, and need for cognition and mood regulation do not play as strong a role as previously thought.
The study suggests that educators should be aware of their level of cognitive absorption and reinforcement
loops when engaging with AI to prevent habitual over-reliance. Furthermore, the absence of direct effects
from mood regulation and positive affect on AI addiction indicates that teachers who enjoy AI use or rely
on it for stress relief are not necessarily at risk of compulsive engagement, reinforcing the idea that AI
should be used as a supportive tool rather than a replacement for professional autonomy.

6.2 Theoretical implications

The findings of this study extend the I-PACE model (Brand et al., 2016, 2019) by demonstrating its
applicability to the domain of generative AI addiction among educators. While previous research has
primarily applied the model to internet gaming disorder and social media addiction (Lörsch et al., 2025;
Pupi et al., 2024), this study confirms that cognitive absorption plays a central role in sustaining AI
dependency among teachers. Unlike passive digital consumption, AI engagement involves interactive
problem-solving, knowledge generation, and decision-making, making its addictive potential distinct (Suva
& Bhatia, 2024). The study highlights how perceived usefulness, self-efficacy, and mood regulation
indirectly contribute to AI addiction through cognitive absorption, reinforcing the importance of
psychological and behavioral factors in shaping compulsive AI use. These insights support the theoretical
expansion of the I-PACE framework by demonstrating how specific cognitive and emotional mechanisms
interact to foster AI dependence in professional settings, particularly within educational environments
where generative AI is integrated into daily teaching practices (Bower et al., 2024).

Furthermore, the study challenges existing literature regarding the roles of need for cognition and mood
regulation in technology addiction. Previous research suggested that individuals with high need for
cognition are more likely to engage in prolonged technology use due to intellectual curiosity (Tanas et al.,
2020; Cacioppo & Petty, 1982). However, this study finds no direct link between need for cognition and AI
addiction, indicating that teachers who seek cognitive stimulation from AI do so in a structured and
intentional manner rather than through habitual overuse. Similarly, mood regulation—often associated with
digital addiction as a coping mechanism (Hu et al., 2023; Ye et al., 2024)—does not directly contribute to
AI addiction in this study. This suggests that teachers use AI primarily for functional efficiency rather than
as an emotional escape, reinforcing the need to differentiate between AI reliance for workload management
and compulsive AI engagement. These findings contribute to refining theoretical perspectives on digital
addiction by demonstrating that professional AI use may operate under different psychological mechanisms
compared to entertainment-based digital dependencies.

Finally, this study contributes to the broader discourse on digital well-being in professional settings by
emphasizing the role of cognitive engagement in technology addiction. Unlike previous models of
technology addiction, which have primarily focused on leisure-based technologies such as social media and
gaming (Renshaw & Carley, 2024; Piko et al., 2025), this study highlights that AI dependence in educators
arises from work-related cognitive immersion rather than hedonic motivations. This distinction underscores
the necessity for future research to refine existing addiction models to account for AI’s dual role as both a
productivity tool and a potential source of compulsive engagement. By incorporating AI-specific
mechanisms into addiction frameworks, scholars can develop more nuanced theoretical approaches to
understanding and mitigating AI dependency in education and other professional domains. This research
thus lays the groundwork for more comprehensive models that address the cognitive, affective, and
behavioral dimensions of AI addiction among educators (Zhong et al., 2024).

6.3 Practical implications

The findings of this study have important implications for educators, institutions, and policymakers seeking
to balance the benefits of generative AI with the risks of over-reliance. One key implication is the need for
professional development programs that promote responsible AI use among teachers. Since cognitive
absorption is a major factor in AI addiction, training should focus on fostering critical engagement with AI
rather than passive dependence. Institutions should provide workshops on AI literacy, emphasizing how AI
can serve as a complementary tool rather than a replacement for teacher expertise (Zhai, 2024). Additionally,
educators should be encouraged to use AI strategically—leveraging it for efficiency without allowing it to
diminish their creativity, autonomy, and pedagogical decision-making (Bower et al., 2024). By promoting
mindful AI usage, teachers can integrate AI-driven tools effectively while maintaining their cognitive
engagement and instructional authority.

Another practical implication is the importance of AI usage guidelines and digital well-being policies at
institutional and governmental levels. Given that perceived usefulness significantly contributes to AI
dependence, educational institutions should implement structured policies that set boundaries for AI
integration in teaching (Dehghani & Mashhadi, 2024). Schools and universities can establish AI usage
frameworks that encourage teachers to balance AI assistance with traditional pedagogical practices. For
example, institutions could require educators to periodically review and revise AI-generated content to
ensure originality and relevance, preventing over-reliance on AI-generated lesson plans (Chan & Tsi, 2024).
Moreover, policymakers should collaborate with technology developers to design AI tools that encourage
critical thinking rather than merely automating tasks. This can be achieved by embedding features that
prompt educators to make active decisions instead of accepting AI-generated recommendations passively
(Comiskey, 2024).

Finally, the study highlights the need for psychological and organizational support systems to mitigate AI
dependency risks among teachers. Since AI engagement can lead to cognitive absorption and habitual
reliance, educational institutions should offer resources that help teachers maintain a balanced approach to
AI use. Schools can integrate digital wellness programs that raise awareness of AI overuse and encourage
self-regulation strategies (Almulla et al., 2025). For instance, structured AI usage reflection exercises or
digital detox initiatives could help teachers assess their reliance on AI tools and adjust their habits
accordingly (Galvan & Newman, 2025). Additionally, school administrators should foster a culture where
AI is viewed as a tool to enhance, rather than replace, human judgment and professional expertise. By
addressing both the benefits and risks of AI in education, institutions can ensure that teachers maximize
AI's potential while safeguarding their autonomy, creativity, and professional well-being.

6.4 Limitations and future directions

Despite its contributions, this study has several limitations that should be acknowledged. First, the reliance
on self-reported data may introduce biases, as participants might overestimate or underestimate their
dependence on generative AI. While validated measurement scales were used to ensure accuracy, subjective
perceptions can still impact the findings. Future research could incorporate objective data collection
methods, such as AI usage logs or observational studies, to gain a more precise understanding of how
teachers engage with AI tools. Additionally, this study was conducted within a specific educational and
cultural context, which may limit the generalizability of the results. Further research across different
educational systems and regions could help determine whether similar patterns of AI dependence exist in
varied teaching environments.

Another limitation of this study is its cross-sectional design, which captures AI use and dependence at a
single point in time. While the findings suggest relationships between cognitive absorption, perceived
usefulness, and AI addiction, they do not establish causality. Longitudinal studies tracking AI engagement
over time would provide deeper insights into how AI dependence develops and whether certain
psychological or environmental factors contribute to its persistence. Additionally, experimental studies
could explore interventions designed to mitigate AI addiction, such as structured AI training programs,
digital well-being initiatives, or strategies for promoting mindful AI use in educational settings.

Finally, this study primarily focuses on individual psychological and behavioral factors influencing AI
addiction, without extensively examining institutional and technological influences. The role of workplace
policies, school leadership, and AI tool design in shaping teachers’ AI engagement remains an important
area for future exploration. Additionally, as AI continues to evolve, new features such as adaptive learning,
automated decision-making, and personalized recommendations may further impact teachers’ reliance on
AI. Future research should examine how these advancements affect cognitive absorption and whether they
introduce new risks or benefits in educational contexts. Addressing these gaps will provide a more
comprehensive understanding of AI dependence and help develop strategies to ensure responsible AI
integration in teaching.

7. Conclusion
This study examined the factors influencing teachers’ dependence on generative AI through the lens of the
I-PACE model, highlighting cognitive absorption as a central driver of AI addiction. The findings suggest
that while AI tools offer significant benefits in enhancing efficiency and reducing workload, excessive
engagement can lead to habitual reliance, potentially diminishing teachers' autonomy and creativity.
Perceived usefulness, self-efficacy, and mood regulation were found to indirectly contribute to AI
dependence, reinforcing the importance of mindful AI integration in educational practices. Additionally, the
study challenges assumptions about the role of need for cognition and mood regulation in technology
addiction, suggesting that AI reliance among educators is primarily driven by cognitive engagement rather
than emotional dependence. These insights have practical implications for educators, institutions, and
policymakers in designing strategies to promote responsible AI use, balancing its advantages while
mitigating risks of over-reliance. Future research should explore long-term AI engagement trends,
institutional policies, and emerging AI features to ensure that AI remains a supportive tool rather than a
disruptive force in education.

References
Acosta-Enriquez, B. G., Ballesteros, M. A. A., Guzman Valle, M. D. L. A., Angaspilco, J. E. M., Aquino
Lalupú, J. D. R., Jaico, J. L. B., Germán Reyes, N. C., Alarcón García, R. E., & Castillo, W. E. J.
(2025). The mediating role of academic stress, critical thinking and performance expectations in
the influence of academic self-efficacy on AI dependence: Case study in college students.
Computers and Education: Artificial Intelligence, 8, 100381.
[Link]
Agarwal, R., & Karahanna, E. (2000). Time Flies When You’re Having Fun: Cognitive Absorption and
Beliefs about Information Technology Usage. MIS Quarterly, 24(4), 665.
[Link]
Almulla, M. O., Alismail, A. M., Mahama, I., & Almulla, A. A. (2025). Resilience as a predictor of internet
addictive behaviours: A study among Ghanaian and Saudi samples using structural equation
modelling approach. BMC Psychology, 13(1), 77. [Link]
Baı̇ Doo-Anu, D., & Owusu Ansah, L. (2023). Education in the Era of Generative Artificial Intelligence
(AI): Understanding the Potential Benefits of ChatGPT in Promoting Teaching and Learning.
Journal of AI, 7(1), 52–62. [Link]
Bewersdorff, A., Hornberger, M., Nerdel, C., & Schiff, D. S. (2025). AI advocates and cautious critics: How
AI attitudes, AI interest, use of AI, and AI literacy build university students’ AI self-efficacy.
Computers and Education: Artificial Intelligence, 8, 100340.
[Link]
Bower, M., Torrington, J., Lai, J. W. M., Petocz, P., & Alfano, M. (2024). How should we change teaching
and assessment in response to increasingly powerful generative Artificial Intelligence? Outcomes
of the ChatGPT teacher survey. Education and Information Technologies.
[Link]
Bozoglan, B., Demirer, V., & Sahin, I. (2014). Problematic Internet use: Functions of use, cognitive
absorption, and depression. Computers in Human Behavior, 37, 117–123.
[Link]
Brand, M., Wegmann, E., Stark, R., Müller, A., Wölfling, K., Robbins, T. W., & Potenza, M. N. (2019). The
Interaction of Person-Affect-Cognition-Execution (I-PACE) model for addictive behaviors: Update,
generalization to addictive behaviors beyond internet-use disorders, and specification of the process
character of addictive behaviors. Neuroscience & Biobehavioral Reviews, 104, 1–10.
[Link]
Brand, M., Young, K. S., Laier, C., Wölfling, K., & Potenza, M. N. (2016). Integrating psychological and
neurobiological considerations regarding the development and maintenance of specific Internet-
use disorders: An Interaction of Person-Affect-Cognition-Execution (I-PACE) model.
Neuroscience & Biobehavioral Reviews, 71, 252–266.
[Link]
Cacioppo, J. T., & Petty, R. E. (1982). The need for cognition. Journal of Personality and Social Psychology,
42(1), 116–131. [Link]
Calaresi, D., Cuzzocrea, F., Saladino, V., & Verrastro, V. (2024). The Relationship Between Trait Emotional
Intelligence and Problematic Social Media Use. Youth & Society, 0044118X241273407.
[Link]
Celik, I., Dindar, M., Muukkonen, H., & Järvelä, S. (2022). The Promises and Challenges of Artificial
Intelligence for Teachers: A Systematic Review of Research. TechTrends, 66(4), 616–630.
[Link]
Chan, C. K. Y., & Tsi, L. H. Y. (2024). Will generative AI replace teachers in higher education? A study of
teacher and student perceptions. Studies in Educational Evaluation, 83, 101395.
[Link]
Cheung, R., & Vogel, D. (2013). Predicting user acceptance of collaborative technologies: An extension of
the technology acceptance model for e-learning. Computers & Education, 63, 160–175.
[Link]
Chiu, T. K. F. (2024). The impact of Generative AI (GenAI) on practices, policies and research direction in
education: A case of ChatGPT and Midjourney. Interactive Learning Environments, 32(10), 6187–
6203. [Link]
Comiskey, C. (2024). Disruptive Innovation, Sustainability, and Sharing Best Practice: How to Be
Innovative, Sustainable, and Share Learning. In C. Comiskey, Addiction Research and Evaluation
(pp. 185–204). Springer Nature Switzerland. [Link]
Dehghani, H., & Mashhadi, A. (2024). Exploring Iranian english as a foreign language teachers’ acceptance
of ChatGPT in english language teaching: Extending the technology acceptance model. Education
and Information Technologies. [Link]
Dinnie, K. (2024). Immersive, addictive, and professionalized—Emerging trends and future directions in
place branding and public diplomacy. Place Branding and Public Diplomacy, 20(S1), 33–35.
[Link]
Dockray, S., & Steptoe, A. (2010). Positive affect and psychobiological processes. Neuroscience &
Biobehavioral Reviews, 35(1), 69–75. [Link]
Du, Y. (2024). A streamlined approach to scale adaptation: Enhancing validity and feasibility in educational
measurement. Journal of Language Teaching, 4(3), 18–22. [Link]
Fryer, L. K., Thompson, A., Nakao, K., Howarth, M., & Gallacher, A. (2020). Supporting self-efficacy
beliefs and interest as educational inputs and outcomes: Framing AI and Human partnered task
experiences. Learning and Individual Differences, 80, 101850.
[Link]
Galvan, E. G., & Newman, C. L. (2025). EXPRESS: The Digital Detox Paradox: Potential Backfire Effects
of Digital Detox Interventions on Consumer Digital Well-Being. Journal of Public Policy &
Marketing, 07439156251322105. [Link]
Hsu, H., & Lin, C. (2022). Extending the technology acceptance model of college learners’ mobile‐assisted
language learning by incorporating psychological constructs. British Journal of Educational
Technology, 53(2), 286–306. [Link]
Hu, B., Mao, Y., & Kim, K. J. (2023). How social anxiety leads to problematic use of conversational AI:
The roles of loneliness, rumination, and mind perception. Computers in Human Behavior, 145,
107760. [Link]
Hu, L., Wang, H., & Xin, Y. (2025). Factors influencing Chinese pre-service teachers’ adoption of
generative AI in teaching: An empirical study based on UTAUT2 and PLS-SEM. Education and
Information Technologies. [Link]
Huang, F., Wang, Y., & Zhang, H. (2024). Modelling Generative AI Acceptance, Perceived Teachers’
Enthusiasm and Self‐Efficacy to English as a Foreign Language Learners’ Well‐Being in the Digital
Era. European Journal of Education, 59(4), e12770. [Link]
Huang, Y., & Huang, H. (2024). Exploring the Effect of Attachment on Technology Addiction to Generative
AI Chatbots: A Structural Equation Modeling Analysis. International Journal of Human–Computer
Interaction, 1–10. [Link]
Hutchison, J. A., & Gunthert, K. C. (2013). Measuring Beliefs About Automatic Mood Regulation:
Development of a Self-Report Scale. Cognitive Therapy and Research, 37(6), 1243–1256.
[Link]
Joshanloo, M. (2017). Factor structure and criterion validity of original and short versions of the Negative
and Positive Affect Scale (NAPAS). Personality and Individual Differences, 105, 233–237.
[Link]
Kohnke, L., & Moorhouse, B. L. (2025). Enhancing the Emotional Aspects of Language Education through
Generative Artificial Intelligence (GenAI): A Qualitative Investigation. Computers in Human
Behavior, 108600. [Link]
Larsen, R. J. (2000). Toward a Science of Mood Regulation. Psychological Inquiry, 11(3), 129–141.
[Link]
Li, Q., Yu, Y., Wang, X., Wong, S. Y., & Yang, X. (2025). The relationship between parental affective
disorders and digital addiction in children and adolescents: A systematic review and meta-analysis.
Addictive Behaviors, 164, 108282. [Link]
Liang, J., Wang, L., Luo, J., Yan, Y., & Fan, C. (2023). The relationship between student interaction with
generative artificial intelligence and learning achievement: Serial mediating roles of self-efficacy
and cognitive engagement. Frontiers in Psychology, 14, 1285392.
[Link]
Lin, H.-F. (2009). Examination of cognitive absorption influencing the intention to use a virtual community.
Behaviour & Information Technology, 28(5), 421–431.
[Link]
Liu, Y., & Chang, P. (2024). Exploring EFL teachers’ emotional experiences and adaptive expertise in the
context of AI advancements: A positive psychology perspective. System, 126, 103463.
[Link]
Lord, K. R., & Putrevu, S. (2006). Exploring the dimensionality of the need for cognition scale. Psychology
& Marketing, 23(1), 11–34. [Link]
Lörsch, F., Schmid, A. M., Thomas, T. A., Brand, M., Müller, A., & Steins-Loeber, S. (2025). The effect of
individual differences on Pavlovian conditioning in specific Internet-use disorders. Behavioural
Brain Research, 476, 115254. [Link]
Meng, S.-Q., Cheng, J.-L., Li, Y.-Y., Yang, X.-Q., Zheng, J.-W., Chang, X.-W., Shi, Y., Chen, Y., Lu, L.,
Sun, Y., Bao, Y.-P., & Shi, J. (2022). Global prevalence of digital addiction in general population:
A systematic review and meta-analysis. Clinical Psychology Review, 92, 102128.
[Link]
Piko, B. F., Müller, V., Kiss, H., & Mellor, D. (2025). Exploring contributors to FoMO (fear of missing out)
among university students: The role of social comparison, social media addiction, loneliness, and
perfectionism. Acta Psychologica, 253, 104771. [Link]
Pupi, V., Polizzi, S., Mazzoni, E., & Biolcati, R. (2024). #Neuroticgrammers: The mediating role of fear of
missing out, escapism and self-expression motives between neuroticism and problematic Instagram
use. Current Psychology, 43(27), 22831–22842. [Link]
Qu, J., & Wang, Y. (2024). Modelling the interaction between EFL teachers’ personality traits, emotion
regulation strategies, and teachers’ depression: A structural equation approach. European Journal
of Education, 59(4), e12727. [Link]
Renshaw, S. L., & Carley, K. M. (2024). Linking online activity to offline behavior: A meta-review of three
decades of online-to-offline scholarship with future implications for AI. Emerging Trends in Drugs,
Addictions, and Health, 4, 100154. [Link]
Rodríguez-Ruiz, J., Marín-López, I., & Espejo-Siles, R. (2025). Is artificial intelligence use related to self-
control, self-esteem and self-efficacy among university students? Education and Information
Technologies, 30(2), 2507–2524. [Link]
Saadé, R., & Bahli, B. (2005). The impact of cognitive absorption on perceived usefulness and perceived
ease of use in on-line learning: An extension of the technology acceptance model. Information &
Management, 42(2), 317–327. [Link]
Suva, M., & Bhatia, G. (2024). Artificial Intelligence in Addiction: Challenges and Opportunities. Indian
Journal of Psychological Medicine, 02537176241274148.
[Link]
Tanas, L., Winkowska-Nowak, K., & Pobiega, K. (2020). The Importance of Teachers’ Need for Cognition
in Their Use of Technology in Mathematics Instruction. Frontiers in Psychology, 11, 259.
[Link]
Tie, B., Zhang, T., He, M., Geng, L., Feng, Q., Liu, C., Wang, X., Wang, Y., Tian, D., Gao, Y., Wang, P.,
Yang, W., & Qiu, J. (2025). Smartphone and the brain: Stress and self-control mediate the
association between the connectome-based predictive modeling of fMRI brain network and
problematic smartphone use. Computers in Human Behavior, 165, 108531.
[Link]
Trusz, S., & Babel, P. (2016). Interpersonal and intrapersonal expectancies. Routledge.
Wang, S., Wang, H., Jiang, Y., Li, P., & Yang, W. (2023). Understanding students’ participation of intelligent
teaching: An empirical study considering artificial intelligence usefulness, interactive reward,
satisfaction, university support and enjoyment. Interactive Learning Environments, 31(9), 5633–
5649. [Link]
Wang, Y., Cao, Y., Gong, S., Wang, Z., Li, N., & Ai, L. (2022). Interaction and learning engagement in
online learning: The mediating roles of online learning self-efficacy and academic emotions.
Learning and Individual Differences, 94, 102128. [Link]
Wang, Y.-Y., & Chuang, Y.-W. (2024). Artificial intelligence self-efficacy: Scale development and
validation. Education and Information Technologies, 29(4), 4785–4808.
[Link]
Ward, M. K., & Meade, A. W. (2023). Dealing with Careless Responding in Survey Data: Prevention,
Identification, and Recommended Best Practices. Annual Review of Psychology, 74(1), 577–596.
[Link]
Wen, Y., Zhao, X., Li, X., & Zang, Y. (2025). Attitude Mining Toward Generative Artificial Intelligence in
Education: The Challenges and Responses for Sustainable Development in Education.
Sustainability, 17(3), 1127. [Link]
Yao, N., & Wang, Q. (2024). Factors influencing pre-service special education teachers’ intention toward
AI in education: Digital literacy, teacher self-efficacy, perceived ease of use, and perceived
usefulness. Heliyon, 10(14), e34894. [Link]
Zhai, X. (2024). Transforming Teachers’ Roles and Agencies in the Era of Generative AI: Perceptions,
Acceptance, Knowledge, and Practices. Journal of Science Education and Technology.
[Link]
Zhang, C., Schießl, J., Plößl, L., Hofmann, F., & Gläser-Zikuda, M. (2023). Acceptance of artificial
intelligence among pre-service teachers: A multigroup analysis. International Journal of
Educational Technology in Higher Education, 20(1), 49. [Link]
00420-7
Zhang, S., Zhao, X., Zhou, T., & Kim, J. H. (2024). Do you have AI dependency? The roles of academic
self-efficacy, academic stress, and performance expectations on problematic AI usage behavior.
International Journal of Educational Technology in Higher Education, 21(1), 34.
[Link]
Zhong, W., Luo, J., & Lyu, Y. (2024). How Do Personal Attributes Shape AI Dependency in Chinese Higher
Education Context? Insights from Needs Frustration Perspective. PLOS ONE, 19(11), e0313314.
[Link]
Zhu, J., & Fong, L. H. N. (2025). Self-control and problematic short-form video usage: The mediating roles
of automaticity and value-driven attention. Behaviour & Information Technology, 1–11.
[Link]
Zou, B., Du, Y., Wang, Z., Chen, J., & Zhang, W. (2023). An Investigation into Artificial Intelligence speech
evaluation programs with automatic feedback for developing EFL learners’ speaking skills. SAGE
Open, 13(3). [Link]
Zou, B., Liviero, S., Ma, Q., Zhang, W., Du, Y., & Xing, P. (2024). Exploring EFL learners’ perceived
promise and limitations of using an artificial intelligence speech evaluation system for speaking
practice. System, 126, 103497. [Link]