International Journal of Information and Education Technology, Vol. 15, No.

3, 2025

ChatGPT in Education: Investigating Students Online

Learning Behaviors
Anindita Pratita1, Tri Lathif Mardi Suryanto1,*, Arista Pratama1, and Adi Wibowo2
1Department of Information System, Faculty of Computer Science, Universitas Pembangunan Nasional “Veteran” Jawa Timur, Surabaya,
East Java, Indonesia
2Department of Informatics, Faculty of Science and Mathematics, Universitas Diponegoro, Semarang, Indonesia

Email: [email protected] (A.P.); [email protected]; (T.L.M.S.); [email protected] (A.P.);

[email protected] (A.W.)
*Corresponding author

Manuscript received September 22, 2024; revised October 18, 2024; accepted December 5, 2024; published March 17, 2025

Abstract—The research aims to investigate the intentions and scores, these soft skills are crucial for students to develop into
behaviour of high school students regarding their use of lifelong learners who can continue to have successful
ChatGPT. Self-directed learning can be challenging due to a professional lives [1].
lack of formal structure and external motivation. ChatGPT is a
tool that can help address these challenges by providing
Self-learning encounters difficulties sustaining motivation
personalized assistance, immediate feedback, and customized and self-control due to the absence of a formal framework and
educational resources to enhance self-directed learning. The external responsibility in conventional educational
Unified Theory of Acceptance and Use of Technology (UTAUT) environments [2]. This can result in the tendency to delay
framework was adapted to align with the study’s case study tasks and experience emotional volatility, harming one’s
requirements, with 385 respondents from four public high progress. The reliability and trustworthiness of information
schools. The data analysis employed Structural Equation
Modelling (SEM) through a Path Analysis approach. The
sources are of utmost importance, as learners need help
conceptual model was developed, and hypotheses were tested. discerning between accurate and false information, which can
The validity and reliability of the measurement model were lead to misunderstandings and incomplete knowledge [3].
assessed through outer model analysis. The main contribution ChatGPT is a versatile educational aid that is always
of this paper is investigating the factors that influence high available, and it elucidates elucidations of challenging ideas
school students’ intentions and behaviour in using ChatGPT by and supplies supplementary pertinent sources and materials.
adapting the UTAUT framework to the context of ChatGPT
usage in high schools. The study reveals that social influence,
It can enhance motivation and self-discipline by offering
hedonic motivation, and habit significantly influence high school personalized interactions and immediate replies, reinforcing
students’ intention to use ChatGPT. Peer recommendations, sustained learning. Furthermore, ChatGPT can customize
enjoyment, and satisfaction influence these factors. The primary educational resources according to individual requirements,
drivers of actual usage are facilitating conditions, habit, and facilitating the exploration of many subjects and acquiring
behavioural intention. Performance expectancy and Effort essential knowledge [4].
expectancy do not significantly affect students’ intention to use
ChatGPT. The study emphasizes fostering a supportive social
Therefore, the factors influencing students’ self-learning
environment and positive habitual interactions to enhance using ChatGPT should be measured and analyzed. According
ChatGPT adoption and usage. The research results among to Zhao et al., understanding these factors is crucial for
students are designed to be studied in other developing countries tailoring AI tools to enhance the learning experience [5]. Key
so that researchers can examine students’ self-learning aspects to evaluate include the effectiveness of ChatGPT in
motivation through AI. providing accurate and relevant information, the impact of its
interactive features on maintaining motivation and
Keywords—AI in education, ChatGPT, Unified Theory of
Acceptance and Use of Technology (UTAUT), self-directed engagement, and its role in overcoming challenges such as
learning information overload and quality of resources [6]. By
examining these elements, educators and learners can better
I. INTRODUCTION leverage ChatGPT to support self-directed learning, ensuring
Self-learning is learning independently without formal that it is valuable in achieving educational goals and fostering
instruction from a teacher or educational institution. It a more effective and personalized learning environment.
involves using resources such as books, articles, videos, and Recent studies have explored the use of ChatGPT in
digital tools to develop skills and knowledge. Self-learning is Indonesian education, highlighting both benefits and
essential in the digital age to adapt to technological challenges. According to Harunasari, it can be effectively
advancements and industry changes. It allows individuals to integrated into EFL writing classes with clear usage policies
develop new skills and deepen knowledge independently, and proper monitoring [7]. Its usage positively impacted
supporting personal and professional development. It students’ learning motivation, particularly among younger
encourages creativity and innovation by providing the male students [8]. However, using ChatGPT raises legal and
freedom to explore new topics without the limitations of a ethical concerns, potentially leading to copyright
formal education. Self-learning also encourages students to infringement and ethical violations in academic settings.
examine various sources of knowledge independently, Despite these challenges, ChatGPT offers innovative
provides for flexible learning, and fosters the growth of self- educational opportunities, such as creating learning materials,
discipline. Rather than only focusing on raising grades or test semester plans, and textbooks. It can assist educators in
fulfilling their teaching, research, and community service

doi: 10.18178/ijiet.2025.15.3.2262 510

 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

responsibilities [9]. These studies emphasize the need for surveys rating ChatGPT’s utility in educational environments.
responsible implementation of AI tools like ChatGPT in While prospective instructors see ChatGPT as helpful in
Indonesian education, balancing their potential benefits with developing educational materials and facilitating learning,
ethical considerations and proper guidelines. worries persist about its potential to reduce critical thinking
The study was conducted in four public high schools, each and the difficulties of identifying plagiarism. In particular,
with distinct policies and infrastructure supporting AI humanities students indicated more concern about
integration in education. These schools have been proactive information quality and academic integrity than their science-
in adopting AI technologies to enhance learning experiences, focused colleagues.
aligning with broader educational trends [1, 10, 11]. This study contributes to the discussion on AI in education
The study employed a mixed-methods approach using an by emphasizing the significance of designing customized AI
online survey with closed and open-ended questions, integration strategies considering educators’ different
sampling 583 participants (399 students and 184 teachers) experiences. It emphasizes the importance of addressing the
from Hong Kong higher education institutions. Quantitative potential issues that AI tools may provide in sustaining
analysis was conducted through descriptive statistics and t- educational standards and integrity. However, the authors
tests to detect significant differences in Gen AI adoption and recognize that the study’s sample size and demographics are
perceptions between students and teachers, using a 5-point limited, which may affect the generalizability of the results.
Likert scale with an additional “Not Sure” option. The Furthermore, the study did not examine the long-term
qualitative component involved thematic coding of open- consequences of AI integration on teaching practices,
ended responses by two independent coders, achieving indicating a need for future research [14].
intercoder agreement rates between 72% and 77% [12]. The study examines Korean in-service teachers’
Chan et al. note that This study’s limitations highlight the impressions of AI education in K-12 settings, and their
need for cross-cultural comparisons, as its primary focus on experiences with AI training, with the goal of better
Hong Kong restricts the generalizability of the findings. The understanding teachers’ perspectives, which are essential for
reliance on self-reported data introduces potential social effective implementation but have received less attention than
desirability bias, and the lack of longitudinal data limits student viewpoints. Using an online survey of 20 instructors
understanding of Gen AI’s lasting impacts on learning and a focus group with four participants, the study discovered
outcomes. Additionally, basing generational attitudes solely that teachers usually have positive opinions regarding AI in
on professional roles rather than age groups restricts a more education and want to adopt AI-based platforms. Teachers
nuanced generational analysis. expected AI to improve student skills (50%), promote
A recent study investigated the impact of ChatGPT on PhD professional growth (45%), and reduce workload (30%).
students’ research methods, focusing on perceived ease of use, However, opinions on its usefulness in student counseling
usefulness, and satisfaction as factors influencing the tool’s differed.
utilization and academic value. This quantitative study, which This study emphasizes the necessity of practical and hands-
included 80 PhD students from Morocco, analyzed data using on AI training in teacher development. While AI training
a Likert-scale questionnaire and Partial Least Squares enhances knowledge and professional progress, it currently
Structural Equation Modeling (PLS-SEM) to ensure needs more practical use due to its theoretical focus. The
reliability and validity. The findings demonstrated that authors argue for training programs that incorporate project-
perceived ease of use positively impacts perceived usefulness based applications matched with curricula, arguing that
and satisfaction but does not directly improve ChatGPT usage. better-designed training can improve educational outcomes
The study concludes that it can appropriately boost PhD and assist teachers in integrating AI. The paper also
students’ productivity and contentment, resulting in better recommends topics for future research to improve AI training
research results. and close gaps identified by educators [15].
The study highlights the potential of AI tools such as The practical consequences of the research findings on
ChatGPT in academic settings, advocating for their use as students’ use of ChatGPT should be sufficiently covered in
supportive resources rather than replacements for human the text. While the results show which factors affect students’
work as written by Boubker, Omar et al. The study intents and behaviors while using ChatGPT, more
emphasizes the importance of both skill development and information must be provided to understand how these
ethical issues in AI use by suggesting the incorporation of AI findings might be used in regular classroom settings. A more
training into PhD programs. Despite its significant insights, thorough explanation is required to demonstrate how
the study admits limitations, such as emphasizing a Moroccan ChatGPT may improve learning effectiveness, foster
student sample and using only quantitative methodologies. responsible technology usage habits, and address ethical and
Future studies could benefit from mixed-method approaches privacy concerns. Also, the significance of having suitable
and an examination of institutional support to acquire a more infrastructure and access for ChatGPT must be sufficiently
complete picture of AI’s function in PhD education [13]. discussed. Given the need to provide all students with fair
The article “ChatGPT’s Role in the Education System: access to technology, the document should analyze the
Insights from the Research” investigates the perceived requirements for digital infrastructure, including dependable
benefits and problems of ChatGPT among future educators, internet connections and sufficient gadgets. The advantages
emphasizing its ability to improve instruction and its impact of utilizing ChatGPT in the classroom might become unequal
on academic integrity and critical thinking. The research, and restricted to students with solid access without the
which included 70 Master’s students in Teacher Training, necessary infrastructure.
collected data from science and humanities backgrounds via Implementing a learning model that combines ChatGPT

511
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

with conventional and digital teaching techniques is one mimic human writing [4]. As a chatbot, ChatGPT can respond
innovative addition that may be made. Case studies or pilot to text-based user queries, complete tasks, and address
programs at several schools to evaluate the efficacy of these productivity challenges [18]. AI advancements are
methods might be a part of this. This study’s results might revolutionizing the education sector, enabling students to
significantly impact education by fusing AI technology with gain knowledge and access information through AI
current teaching methods. While there are numerous studies applications [19].
on adoption models and AI, this study distinguishes itself by Educators view ChatGPT as a potential tool to transform
focusing on the context of high school students’ use of and enhance future learning and research. However, there are
ChatGPT for learning, which has yet to be extensively concerns among some educators about its ability to generate
explored. The selection of particular constructs Performance content that should be used more by students. ChatGPT can
Expectancy (PE), Effort Expectancy (EE), Social Influence engage in humanlike conversations and develop content
(SI), Facilitating Conditions (FC), Hedonic Motivation (HM), based on user input. The rapid growth of artificial intelligence
Habit (H), Behavioral Intention (BI), and ChatGPT Use (AI) is creating new habits, with society increasingly relying
(GPTU) is grounded in the Unified Theory of Acceptance and on AI to simplify daily activities. This rapid growth raises
Use of Technology (UTAUT) model. concerns due to the need for a more adequate understanding
of AI usage. Addressing this issue requires a better
II. LITERATURE REVIEW understanding of AI functions within society. In the academic
realm, AI’s development threatens to replace creative roles
A. UTAUT
traditionally held by students, such as writing scientific and
The Unified Theory of Acceptance and Use of Technology academic papers. Researchers have developed chatbots, AI,
(UTAUT) is a model for technology acceptance and usage and ChatGPT to offer various conversational services tailored
proposed by Venkatesh et al. This model includes four key to human needs [20]
variables: performance expectancy, effort expectancy, social AI in education significantly impacts students by making
influence, and facilitating conditions. Performance their tasks and writing processes quicker and more efficient.
expectancy relates to five variables from previous models: AI has been widely implemented in educational settings,
perceived usefulness (TAM, TAM2, and CTAMTPB), including schools and higher education institutions. ChatGPT
extrinsic motivation (MM), job fit (MPCU), relative and AI applications can improve the efficiency of creating
advantage (IDT), and outcome expectations (SCT). Effort assignments, essays, and papers. However, their use also
expectancy corresponds to three variables from previous poses risks related to plagiarism, and there are concerns that
models: perceived ease of use (TAM, TAM2), complexity students may lose creativity and critical thinking skills by
(MPCU), and ease of use (IDT). Social influence relates to relying too heavily on this technology. Therefore, this
four variables from previous models: subjective norm (TRA, research aims to investigate the intentions and behavior of
TAM2, TPM/DTPB, and C-TAM-TPB), social factors high school students regarding their use of ChatGPT.
(MPCU), and image (TAM2, IDT). Facilitating conditions Students accessed ChatGPT through school-provided
relate to three variables from previous models: perceived devices, including computers and tablets, available in
behavioral control (TPB, DTPB, CTAMTPB), facilitating computer labs and classrooms. The schools ensured these
conditions (MPCU), and compatibility (IDT) [16]. devices had stable internet connections to facilitate the
The UTAUT method is a user acceptance research model seamless use of AI tools. Training sessions and tutorials were
that explains user intentions to use a system and subsequent also provided to students and teachers to familiarize them
usage behavior. According to Venkatesh et al., the strength with ChatGPT and its functionalities.
of UTAUT lies in its ability to explain how individual The schools have implemented specific policies to regulate
differences influence technology usage, elucidating the AI technologies, including ChatGPT. Considering the age
relationships between perceived benefits, ease of use, and the restrictions and potential risks associated with AI, these
intention to use technology [16]. In 2012, Venkatesh and policies are designed to ensure safe and ethical use. Critical
colleagues developed the UTAUT model further into aspects of these policies include [4]:
UTAUT 2. The focus of UTAUT 2 is on the user context, 1) Age Restrictions: Given that ChatGPT has an age
adding new variables such as habit, hedonic motivation, and restriction, the schools obtained parental consent for
price value. For analyzing technology acceptance and usage, students under 18 to use the tool. This was part of a
UTAUT 2 is more appropriate as it is the latest model, broader effort to comply with legal and ethical standards.
incorporating and summarizing eight existing technology 2) Usage Guidelines: The schools established clear
acceptance theories. UTAUT 2 provides insights into guidelines on how ChatGPT should be used in educational
technology acceptance within the context of consumer settings. These guidelines emphasize the importance of
use [17]. using the tool for educational purposes only and
B. ChatGPT in Education discourage misuse, such as generating content for
The swift advancement of technology has dramatically assignments without proper understanding.
facilitated daily human activities. One notable technological 3) Privacy and Security: Policies were implemented to
innovation is ChatGPT, which aims to assist in efficiently protect students’ privacy and data security. This includes
finding information and generating written content. measures to ensure that any data generated or shared
Introduced by OpenAI in November 2022, the Generative through ChatGPT is handled in compliance with data
Pre-Trained Transformer (GPT) is a form of artificial protection regulations.
intelligence capable of producing text responses that closely 4) Monitoring and Evaluation: The schools regularly

512
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

monitor the use of ChatGPT to ensure compliance with technology [14]. Hedonic motivation has a role in
the established policies. Feedback from students and determining technology acceptance in research conducted by
teachers is collected to continuously improve the Alhwaiti [21]. Then, the researcher wants to find out whether
implementation and effectiveness of AI tools in education. using the ChatGPT application for students can cause a sense
of pleasure and satisfaction. The ChatGPT application is
C. Hypotheses
currently easily accessible to anyone via the Internet so that
1) Performance Expectancy (PE) students become happy when using it. Previous studies have
According to Venkatesh et al., performance expectancy is shown that hedonic motivation significantly predicted
the extent to which a person believes using the system will behavioral intention. Hedonic motivation positively and
help him achieve his goals [16]. Previous studies from substantially affects behavioral intention to use ChatGPT in
Alhwaiti, Wu et al. stated that PE was a significant predictor higher education. Hedonic motivation positively and
of BI. The study found that performance expectancy is the substantially impacts behavioral intention to use ChatGPT for
most vital variable in determining a person’s behavior when English language learning [30, 31]. So, the following
adopting technology. One of its applications can be seen from hypothesis is proposed:
the ChatGPT application feature, which makes it easy for H4: Hedonic Motivation (HM) positively and significantly
users to find answers to solve student problems and complete affects Behavioral Intention (BI).
student assignments [21–23]. The convenience is then felt by 5) Facilitating Conditions (FC)
users so that it can influence their behavior when using
According to Venkatesh et al., facilitating conditions is the
artificial intelligence applications. Thus, the following
extent to which a person believes that organizational and
hypothesis is proposed:
technical infrastructure exists to support system use [16].
H1: Performance Expectancy (PE) positively and
Research by García-Murillo et al. shows that facilitating
significantly affects Behavioral Intention (BI).
conditions influence behavioral intention [32]. The use of
2) Effort Expectancy (EE) technology may not run when there are no supporting
Effort Expectancy, according to Venkatesh et al., is how facilities, even if there is an intention or desire for and support
much effort the ease of use is associated with using the from the school environment to use the technology.
system [13]. This variable is identical to the ease of use Researchers want to know how students use the ChatGPT
construct described in research by Alhwaiti [21]. Effort application. Previous research has shown that Facilitating
expectancy (EE) was shown to have an indirect effect on conditions have a positive and significant effect on behavioral
Behavioral Intention through other factors [10, 24]. Students’ intention ChatGPT among higher education students in the
use of the ChatGPT application depends on individual Philippines [33, 34]. Thus, the following hypothesis is
behavioral characteristics. How often students’ experience proposed:
using technology can also affect the ease of students in using H5: Facilitating Conditions (FC) positively and
the ChatGPT application. The main findings of the study significantly affects Behavioral Intention (BI).
from Chen et al., that factors such effort expectancy were Based on the explanation of the Facilitating Conditions
positively associated with students’ behavioral intention to variable, it is then connected to the context of the ChatGPT
use AI for language learning [25]. So, the following application for public high school students. Suppose students
hypothesis is proposed: feel that the conditions that facilitate using the ChatGPT
H2: Effort Expectancy (EE) positively and significantly application already exist and are supported by the school. In
affects Behavioral Intention (BI). that case, students tend actually to intend to use the
3) Social Influence (SI) application. Furthermore, in research by Romero-Rodrí guez
According to Venkatesh et al., social influence is the extent et al., facilitating conditions also positively impact using the
to which a person perceives that other people believe he ChatGPT application (GPTU) [35]. The study found that
should use a new system [16]. The views of teachers and close facilitating conditions positively and significantly affect
relatives in the school environment can influence student students’ behavioral intention to use ChatGPT [36]. Further
behavior toward using artificial intelligence applications research can be conducted to collect data from students
because the teacher influences his students to use the evaluating the extent to which school facilities can support
application. Previous research by Habibi et al. found that using the ChatGPT application. Thus, the following
social influence significantly affects behavioral intention [26]. hypothesis is proposed:
Among the most potent predictors of students’ behavioral H6: Facilitating Conditions (FC) positively and
intention, prior research has shown that studying AI is helpful significantly affects ChatGPT Use (GPTU).
for society, and social influence significantly influences 6) Habit (H)
behavioral intentions to utilize AI in education [27, 28]. According to Venkatesh et al., Habit is the extent to which
Social influence positively and significantly impacts a person tends to use technology automatically due to prior
behavioral intention toward online learning during COVID- learning with the habit of using technology as an
19 [29]. Thus, the following hypothesis is proposed: indicator [17]. The effect of habit variables as predictors of
H3: Social Influence (SI) positively and significantly usage intention has been analyzed in several studies. The
affects Behavioral Intention (BI). user’s willingness to use the application continuously relates
4) Hedonic Motivation (HM) to the user’s habits while using the service [26]. Based on
According to Venkatesh et al., Hedonic motivation is Strzelecki, this study developed and tested a model to
satisfaction or pleasure obtained by a person through examine the predictors of adoption and use of ChatGPT

513
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

among higher education students. Habit has a positive and problems, and theories of technology acceptance solutions
significant effect on behavioral intention to use ChatGPT in that have been presented provide a hypothesis of the model
higher education [31]. Habit can be formed when users that will be used in measurement in this study.
repeatedly or continuously use the ChatGPT application at
school. So, the following hypothesis is proposed:
H7: Habit (H) positively and significantly affects
Behavioral Intention (BI).
Based on the explanation of the Habit variable, it is then
connected to the context of the ChatGPT application for
students. Habit is the tendency of high school students to use
the ChatGPT application automatically, which may have
been formed from repeated or continuous use in the school
environment. If students have a habit of using the ChatGPT
application, then this will have a positive effect on student’s
intention to use the ChatGPT application. In research, Habibi Fig. 1. Proposed technology adoption measurement model.
et al. stated that the habit of using the ChatGPT application
will also positively affect the actual level of use of the This model has been widely validated in various contexts
application [26]. Recent studies have explored the acceptance but is particularly relevant as it provides a comprehensive
and use of ChatGPT in higher education using the UTAUT framework for understanding technology adoption in
model. This study examined the factors influencing college educational settings. The constructs were chosen based on
students’ satisfaction and continued use of the AI chatbot their proven significance in predicting technology adoption
ChatGPT in higher education. The results show that Habit behaviors. For instance, Performance Expectancy is a critical
positively affects the continued use intention of ChatGPT in predictor of Behavioral Intention, as it reflects the degree to
higher education [37]. Thus, the following hypothesis is which students believe that using ChatGPT will help them
proposed: achieve their learning goals. Similarly, Social Influence and
H8: Habit (H) positively and significantly affects ChatGPT Hedonic Motivation are essential in driving actual usage,
Use (GPTU). emphasizing the role of social dynamics and intrinsic
7) Behavioral Intention (BI) enjoyment in technology adoption. Moreover, this study
According to Venkatesh et al., behavioral intention is the incorporates recent findings and extends the UTAUT model
extent to which someone wants to use technology in the by including constructs like Habit, which has shown to be
future [16]. Research conducted by Habibi et al. states that a influential in the context of AI tools.
person’s intention to use a system has a significant effect [26]. This approach allows for a more nuanced understanding of
This research is carried out to determine what variables the factors that drive high school students’ adoption of
influence the behavior of public high school students using ChatGPT, providing insights that can inform targeted
the ChatGPT application. Recent studies have explored the interventions and support mechanisms in educational
factors affecting students’ acceptance and use of ChatGPT in institutions. In summary, this research contributes to the
higher education. These studies consistently found that existing literature by applying a well-established theoretical
Behavioral Intention significantly predicts ChatGPT Use. model to a novel context, thereby offering new insights into
The behavioral intention has a positive and significant effect the factors influencing the adoption of AI tools in education.
on ChatGPT use by students [11, 38-40]. So, the following This approach validates the selected constructs’ relevance
hypothesis is proposed: and underscores their importance in enhancing the learning
H9: Behavioral Intention (BI) has a positive and significant experience through AI. The flow of this research starts from
effect on ChatGPT Use (GPTU). observation and literature study, then identifying problems,
developing a conceptual model adopting research by Habibi
III. MATERIALS AND METHODS et al., and developing hypotheses [26]. To determine the
population and research sample and prepare the instruments.
The flow of this research starts from observation and
After that, an instrument test was carried out to determine
literature study, then identifying problems, developing a
whether the instrument used in the questionnaire was valid
conceptual model adopting research by Habibi et al., and
and reliable. The study significantly contributes to
developing hypotheses [26]. To determine the population and
understanding the UTAUT framework, which is used to
research sample and prepare the instruments. After that, an
understand the factors influencing high school students’ use
instrument test was carried out to determine whether the
of ChatGPT for learning. The research reveals that Social
instrument used in the questionnaire was valid and reliable.
Influence, Hedonic Motivation, and Habit are critical in
Fig. 1 is a conceptual model which contains eight variables
driving usage.
used in this research. The model was adopted from Habibi et
al., using the UTAUT 2 method. There are Performance A. Data Collection
Expectancy (PE), Effort Expectancy (EE), Social Influence This study explores the factors influencing students’
(SI), Facilitating Conditions (FC), Hedonic Motivation (HM), behavioral intention to use ChatGPT, an AI-based tool, in
Habit (H), Behavioral Intention (BI), and ChatGPT Use educational settings. The context is rooted in the increasing
(GPTU). This research develops similar studies related to integration of AI technologies in education, which can
using information technology in learning. The background, potentially enhance learning experiences and outcomes.

514
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

Specifically, the study investigates how various constructs 212 female students and 146 male students. Each school has
from the Unified Theory of Acceptance and Use of an equal distribution, 90 students from Senior High School 1
Technology (UTAUT) model, including Performance and Senior High School 2, while 89 students from Senior
Expectancy, Effort Expectancy, Social Influence, and High School 3 and Senior High School 4. The intensity of
Facilitating Conditions, affect students’ intentions to use ChatGPT use by students is also known. A total of 152
ChatGPT. students rarely use it, 46 students use it one week 1 time, 71
Through a Google form survey, respondents choose one of students use it one week 2 times, 63 students use it one week
the alternative answers that have been provided. The more than 2 times, and the remaining 26 students claim to use
population in this study were public high school students in it daily. Based on student interviews on the use of ChatGPT
Pasuruan City who used the ChatGPT application with a show that there is variation in its acceptance and utilization in
population of 3,405. Who use the ChatGPT application and the daily learning process. ChatGPT helped them complete
have a total population of 3,405 students. The sample for this assignments more efficiently and improved the quality of
study was selected using a simple random sampling method. their assignment results. Students felt comfortable using
This research uses probability sampling techniques of simple ChatGPT and considered the tool a valuable addition to
random sampling type to ensure representation across existing learning materials. However, some students
different demographics, including age, gender, and academic expressed difficulty accessing ChatGPT due to infrastructure
discipline. This method was chosen to minimize selection limitations such as unstable internet connection and
bias and ensure that the sample accurately reflects the broader inadequate devices. In addition, some students felt that
student population. Furthermore, the minimum number is ChatGPT sometimes made them too dependent on this
calculated using the Slovin Formula. Furthermore, technology, thus reducing their ability to think critically and
calculating the minimum number required using the Slovin independently. In addition, the interviews also revealed that
𝑁
Formula as below: 𝑛 1+𝑁 .𝑒 2 although ChatGPT can provide more in depth and quicker
Description: explanations, students still need guidance in using this tool
n = Number of samples optimally. Many students felt they needed more training and
N = Total population advice on effectively utilizing ChatGPT in their learning
e = Error tolerance limit (margin of error) 5% process. Some students also highlighted the importance of
Based on the Slovin formula, The population consisted of integrating ChatGPT with traditional learning methods to get
students from four public high schools, with 385 respondents more balanced and in depth results. These students’
participating. experiences of using ChatGPT colored their perceptions of
the effectiveness and usefulness of this technology in
B. Likert Scale supporting their learning.
The Likert scale measures the respondents’ attitudes on the
D. Dataset Analysis Method
questionnaire sheet. Rensis Likert developed it based on a
book written by Hermawan and Amirullah [41]. Five SEM, or Structural Equation Modeling, can be used to
categories or ranks indicate the degree of agreement or answer various research questions in a structured and
disagreement with each statement relating to the assessed comprehensive series of analyses. According to Hair et al.,
object. The scale used in this thesis, namely a scale of 1-5 SEM represents a second-generation model of multivariate
where number 1 represents the statement “Strongly disagree,” analysis techniques. This model allows researchers to test
number 2 represents the statement “Disagree,” number 3 complex relationships between variables, whether they are
represents the statement “Neutral,” number 4 represents the recursive or non-recursive, aiming to provide a
statement “Agree.” Number 5 represents the statement comprehensive picture of a model [42]. According to Gaston,
“Strongly Agree.” as cited in a book by Yamin & Kurniawan, Partial Least
Squares (PLS) can be utilized for confirmatory purposes,
C. Data Presentation such as hypothesis testing and exploratory purposes.
The demographic data of respondents in this study Although PLS tends to be used more as an exploratory tool
consisted of gender, school origin, and intensity of ChatGPT than an affirmative one, it can also predict whether
use. The number of respondents was 358 students with the relationships between variables exist and then formulate
characteristics in Table 1 below: propositions for further testing. Its main objective is to
explain the interrelationships between constructs and to
Table 1. Respondent demographics
emphasize a deep understanding of the value of these
Characteristics Group Qty %
Female 212 59.2 relationships. In this context, it is crucial to consider the
Gender presence of a theory that provides a basis for describing the
Male 146 40.8
SHS 1 90 25.1 model, the selection of variables, the analysis methods
SHS 2 90 25.1 applied, and the interpretation of results [43].
School Origin
SHS 3 89 24.9
SHS 4 89 24.9 The benefits of using SEM for researchers include the
Rarely Use 152 42.4 ability to explain complex variable relationships and the
One week 1 time 46 12.8 direct or indirect effects of one or several variables on others.
ChatGPT Usage
One week 2 times 71 19.8 SEM also offers higher flexibility for researchers to connect
Intensity
One week more than 2
63 17.6 theory with data. Meanwhile, according to Petter & Hadavi,
times
several benefits come with using partial least squares (PLS)
Based on Table 1, it is known that 358 students consist of as a composite based structural equation modeling (SEM)

515
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

method. PLS frequently involves fewer assumptions than variables include Performance Expectancy, Effort
factor-based SEM approaches, allowing researchers to Expectancy, Facilitating Conditions, Social Influence,
concurrently investigate the measurement and structural ChatGPT Use, Behavioral Intention, Hedonic Motivation,
models. PLS is quite powerful for researchers who want to and Habit. Each variable offers unique insights into the
assess a research model using an SEM-based methodology factors influencing individuals’ adoption and use of ChatGPT.
for all of the reasons mentioned above. But, PLS’s immense These include expectations of performance and desired
power also carries a great deal of responsibility [44]. outcomes, supportive conditions for technology use, social
influences, and habits formed through repeated use. This
E. Variable Construct
article will explore these variables to understand how they
In examining modern technologies such as ChatGPT, contribute to the acceptance and usage of ChatGPT. Variables
several vital variables are crucial in understanding user and questionnaire items can be reviewed in Table 2.
behavior and intention to use these technologies. These
Table 2. Variables and questionnaire items
Code Variable Construct Items Questionnaire Items Sources
PE1 Using ChatGPT will help me complete my assignments.
PE2 Using ChatGPT will increase my efficiency in completing assignments.
PE Performance Expectancy PE3 Using ChatGPT will improve the quality of my assignments. [17, 45]
PE4 Using ChatGPT will make me feel more comfortable when working on
assignments.
EE1 I find it easy to access the ChatGPT page.
EE2 I find it easy to interact with ChatGPT via mobile devices.
EE Effort Expectancy [17, 45]
EE3 I find ChatGPT easy to use.
EE4 I find it easy to become proficient in using ChatGPT.
I have the necessary resources (laptop, internet connection, mobile device, desktop,
FC1
etc.) to use ChatGPT.
FC Facilitating Conditions FC2 I have enough knowledge to use ChatGPT. [17, 45]
FC3 My operating system (Windows et al.) is compatible with ChatGPT.
FC4 I can get help from others if I experience difficulties using ChatGPT.
SI1 My friends think that I should use ChatGPT.
SI Social Influence SI2 My family thinks that I should use ChatGPT. [17, 45]
SI3 People who influence my behavior also use ChatGPT.
GPTU1 I frequently use ChatGPT.
GPTU ChatGPT Use GPTU2 I rely on ChatGPT for studying. [46–48]
GPTU3 I use all the functions in ChatGPT.
HM1 I feel happy using ChatGPT.
HM Hedonic Motivation HM2 I feel comfortable using ChatGPT. [46–48]
HM3 I feel entertained using ChatGPT.
BI1 I intend to use ChatGPT during learning activities.
BI2 I enjoy using ChatGPT.
BI Behavioral Intention [17, 45]
BI3 I plan to continue using ChatGPT.
BI4 I recommend that my friends use ChatGPT.
H1 I feel accustomed to using ChatGPT.
H2 I don't think twice before using ChatGPT.
H Habit H3 I feel that using ChatGPT in learning is natural. [17, 49, 50]
H4 I feel spontaneous using ChatGPT in learning.
H5 I am more likely to use ChatGPT for research or school assignments.

Fig. 2. Outer loadings.

516
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

criterion values can also assess discriminant validity. It can

IV. RESULT AND DISCUSSION be seen that the top value in the Fornell-Lecker Criterion is
Furthermore, the instruments that have been prepared must the highest in its column. The Fornell-Lecker values can be
be tested for validity and reliability. At this stage, it is used to found in Table 4.
measure the outer and inner models and hypothesis testing on
the data obtained from as many as 358 students who have Table 3. Validity and reliability test results
Cronbach’s Composite Outer
used ChatGPT. This analysis process is carried out using Code
Alpha
AVE
Reliability
Indicator
Loadings
SmartPLS software version 3.2.9. PE1 0.818
PE2 0.818
A. Result PE 0.833 0.666 0.888
PE3 0.825
1) Outer model PE4 0.831
EE1 0.799
Outer model analysis is carried out to assess the validity EE2 0.816
EE 0.835 0.667 0.889
and reliability of the model, which is used to define the EE3 0.812
relationship between latent variables and their indicators. At EE4 0.838
FC1 0.775
this stage of the analysis, the validity and reliability of the FC2 0.813
model used are tested, including determining the relationship FC 0.774 0.596 0.855
FC3 0.793
between latent variables and their indicators. As in Fig. 2, all FC4 0.702
indicators have met the requirements to be considered valid, SI1 0.880
SI 0.777 0.692 0.870 SI2 0.858
namely the value of outer loadings > 0.7. Table 3 contains all
SI3 0.753
outer loadings values, AVE (Average Variance Extracted) GPTU1 0.874
values, and Composite Reliability. The outer loadings of all GPTU 0.829 0.746 0.898 GPTU2 0.841
indicators have shown > 0.7 so that all indicators can be GPTU3 0.875
HM1 0.883
considered valid. Then, to measure convergent validity, it can
HM 0.845 0.764 0.907 HM2 0.876
be reviewed through the AVE value, where the AVE value HM3 0.863
must be > 0.5 and the Composite Reliability value must be> BI1 0.845
0.7. In Table 2, all AVE values are > 0.5, Cronbach’s Alpha > BI2 0.833
BI 0.870 0.719 0.911
BI3 0.856
0.7, and Composite Reliability value > 0.7, so it can be
BI4 0.858
considered valid and reliable overall. H1 0.825
Discriminant validity is assessed through cross-loadings H2 0.753
and the Fornell-Lecker Criterion values. A variable is H 0.883 0.682 0.915 H3 0.823
considered to meet discriminant validity if the cross-loading H4 0.885
H5 0.838
value of each indicator for the variable is the highest in its
row. The square root of the AVE values or the Fornell-Lecker
Table 4. Fornell-Lacker discriminant validity
PE EE SI HM FC H BI GPTU
PE 0.816
EE 0.692 0.817
SI 0.675 0.573 0.832
HM 0.748 0.665 0.682 0.874
FC 0.642 0.758 0.594 0.647 0.772
H 0.72 0.656 0.771 0.75 0.632 0.826
BI 0.702 0.636 0.762 0.753 0.631 0.851 0.848
GPTU 0.656 0.581 0.733 0.718 0.628 0.816 0.831 0.863

Table 4 shows that the Fornell-Larcker Criterion values for influence behavioral intention by 77.4%. The ChatGPT Use
each independent variable are more significant than the variable is 0.741, meaning that the Behavioral Intention
correlations between independent variables in the same variable affects ChatGPT Use by 74.1%.
column. This indicates that the discriminant validity criteria 3) Hypothesis test
using the Fornell-Larcker Criterion approach have been met.
2) Inner model Table 6. Results of path coefficients analysis
Original
At this stage of the analysis, the R square value is used to Code Sample T-Statistics P-Value Influence
determine whether the variable has a direct influence or not. (O)
This can be seen in Table 5. PE ≥ BI 0.035 0.802 0.423 Not Sig.
EE ≥ BI 0.017 0.331 0.741 Not Sig.
Table 5. R square SI ≥ BI 0.189 3.475 0.001 Sig
HM ≥ BI 0.174 2.807 0.005 Sig
Code R Square
FC ≥ BI 0.050 1.079 0.281 Not Sig.
BI 0.774
FC ≥ GPTU 0.118 3.004 0.003 Sig
GPTU 0.741
H ≥ BI 0.507 7.745 0.000 Sig
H ≥ GPTU 0.353 7.064 0.000 Sig
The results are in Table 5. They show several 0.774 for the BI ≥ GPTU 0.456 9.634 0.000 Sig
Behavioral Intention variable, meaning that the Performance
Expectation, Effort Expectation, Social Influence, Hedonic From Table 6. three variables do not have a significant
Motivation, Facilitating Conditions, and Habit variables influence because the T Statistics value < 1.9, namely

517
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

Performance Expectancy (PE), Effort Expectancy (EE), and sample value is positive, namely 0.053. Therefore,
Facilitating Conditions (FC) on Behavioral Intention (BI). Facilitating Conditions (FC) do not significantly affect
Meanwhile, the variables Social Influence (SI), Hedonic Behavioral Intention (BI). This means that the facilities
Motivation (HM), and Habit (H) have a significant influence owned by schools and individuals have yet to support
on Behavioral Intention (BI). Then, the variables Facilitating students using ChatGPT. One of them is the internet to access
Conditions (FC), Habit (H), and Behavioral Intention (BI) ChatGPT. This can be an evaluation of schools to support the
also have a significant influence on ChatGPT Use (GPTU). use of ChatGPT at school by improving internet facilities so
that they can be used for the public and are evenly distributed.
B. Discussion
Therefore, the condition of the facility is not a factor in the
The findings of the SEM PLS analysis indicate that level of use of ChatGPT by State High School students.
ChatGPT can effectively assist high school pupils in A comparison between the findings of this study and
enhancing their learning. Several factors, such as Social existing literature, specifically Habibi’s research, reveals
Influence, Hedonic Motivation, and Habit, impact the extent significant differences, particularly in the role of facilitating
of State High School pupils’ inclination to utilize ChatGPT. conditions (FC) on behavioral intentions to use ChatGPT.
The factors influencing the usage of ChatGPT among State The result shows that students who believe they have
High School students are Behavioral Intention, Habit, and sufficient resources to use ChatGPT do not significantly
Facilitating Conditions. impact students’ behavioral intention because the T Statistics
The analysis and correlation test results show that the first value < 1.9. This indicates that access to resources alone
hypothesis cannot be accepted because the T Statstic value is cannot influence students’ intent. Instead, supportive school
<1.9 but declared positive because the original sample value policies that promote equitable access to technology have a
is positive namely 0.041. Therefore, Performance more substantial effect, enabling students to engage with
Expectancy (PE) does not significantly affect Behavioral ChatGPT independently of their access conditions. This
Intention (BI). This means that the use of ChatGPT in collective encouragement mitigates the direct impact of
learning has yet to achieve optimal benefits or advantages for facilitating conditions on students’ behavioral intentions,
State High School students. Therefore, the level of use of underscoring the role of institutional support in fostering
ChatGPT in high school students is not based on ChatGPT technology adoption.
performance expectations. In contrast, the literature presents a robust relationship
Furthermore, the analysis and correlation test cannot between facilitating conditions and actual ChatGPT usage, as
accept the second hypothesis because the T statistic value is observed in Habibi’s study. This relationship likely arises
<1.9. Still, it is declared positive because the original sample from the free availability of essential infrastructure,
value is positive, namely 0.016. Effort Expectancy (EE) does particularly internet access, which significantly influences
not substantially impact Behavioral Intention (BI). student adoption of ChatGPT. Habibi’s findings align with
Consequently, State High School pupils still need to improve previous studies that identified FC as a strong predictor for
in utilizing ChatGPT. It is necessary to assess the using various technological systems. Furthermore, while both
accessibility of the features and appearance to ensure ease of studies corroborate the role of Habit (H) in influencing
use for pupils. Thus, this will facilitate their future interaction ChatGPT usage, their perspectives on Behavioral Intention
with ChatGPT. Hence, the extent to which State High School (BI) differ. Habibi identifies BI as the most significant
students utilize ChatGPT is not determined by something predictor of ChatGPT usage in learning, emphasizing the
other than ChatGPT’s effort expectancy. importance of user intention. In this study, however,
Then, in the analysis and correlation test, the third behavioral intention appears to be less directly influenced by
hypothesis can be accepted because the T Statstic value >1.9 available resources, highlighting instead the critical role of
and the original sample value is positive. So, Social Influence school policies in encouraging students’ engagement with
(SI) significantly positively affects Behavioral Intention (BI). technology. These contrasting insights suggest that while
This means that students’ social influence influences the use resource accessibility can indeed impact usage, supportive
of ChatGPT in learning. They tend to use ChatGPT based on school policies and institutional encouragement are essential
recommendations from their surroundings, such as friends, in shaping students’ behavioral intentions toward technology
relatives, and even teachers at school. Therefore, social use in educational contexts.
influence is one of the factors supporting the level of use of On the other hand, in the analysis and correlation test, the
ChatGPT. sixth hypothesis, namely that Facilitating Conditions (FC)
The analysis and correlation test can accept the fourth has a positive effect on ChatGPT Use (GPTU), can be
hypothesis because the T Statistic value >1.9 and the original accepted. This was found during interviews with students.
sample value are positive. So, Hedonic Motivation (HM) has Teachers at school have conducted counseling sessions on the
a significant positive effect on Behavioral Intention (BI). This use of AI applications. Although 42.4% of students admitted
means state high school students feel satisfied using ChatGPT that they rarely used ChatGPT, they were used to using other
to support learning. They are happy with the answers or AI applications. Therefore, facility conditions can be a factor
solutions ChatGPT provides to complete school assignments. supporting the intensity of ChatGPT.
Therefore, satisfaction is one factor supporting the level of The analysis and correlation test can accept the seventh
use of ChatGPT. hypothesis because the T statistic value > 1.9 and the original
Furthermore, the fifth hypothesis cannot be accepted in the sample value is positive. So, Habit (H) positively affects
analysis and correlation test because the T statistic value is < Behavioral Intention (BI). This means that students’ habits
1.9. However, it is declared positive because the original influence the use of ChatGPT in learning. They tend to use

518
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

ChatGPT to help with research or schoolwork, which causes Experts in education and technology recognize ChatGPT’s
them to become accustomed to using it for schoolwork. potential to enhance self-learning and deepen students’
Therefore, habit is one of the factors supporting the level understanding across various subjects. However, they
of use of ChatGPT. The analysis and correlation test can emphasize the importance of addressing information
accept the eighth hypothesis because the T statistic value > accuracy, data privacy, and ethical use. To ensure effective
1.9 and the original sample value is positive. So, Habit (H) and responsible AI integration, schools must overcome
positively affects ChatGPT Use (GPTU). This is due to their challenges like infrastructure limitations and system
familiarity with utilizing ChatGPT as an educational aid in compatibility while also implementing clear policies and
the academic setting. Therefore, the habit can act as a providing comprehensive training for both teachers and
stimulant for the frequency of ChatGPT usage among students.
students at State High School. ChatGPT can be integrated into the curriculum to support
Furthermore, the ninth hypothesis’s last analysis and learning in subjects like English, History, Science, and
correlation test can be accepted because the T Statistic value > Mathematics. For instance, in English, it can assist with essay
1.9 and the original sample value is positive. So, Behavioral writing, idea generation, and grammar refinement, while
Intention (BI) positively affects ChatGPT Use (GPTU). The teachers help students critically assess AI-generated content.
State High School students’ continued usage of ChatGPT is In History, ChatGPT can summarize texts, create study
primarily influenced by social influence, level of satisfaction, questions, and simulate historical debates, enhancing student
and habit. Research has established a strong correlation engagement. In Science and Math, AI tools can simplify
between the satisfaction level with ChatGPT and the extent complex concepts and generate practice problems for
of the positive effect utilized. Additionally, state high school students. Moreover, ChatGPT can facilitate research, report
students accustomed to using ChatGPT will likely increase writing, and presentations in project-based learning, fostering
their frequency or intensity of usage. Hence, the extent of collaboration as students evaluate and refine AI-suggested
ChatGPT utilization is a contributing element that can ideas.
amplify the frequency of ChatGPT usage among high school Integrating AI tools like ChatGPT in education raises
students. essential ethical issues, including data privacy, security, and
Schools need to increase awareness and social support for potential bias in AI outputs. This study highlights concerns
the use of ChatGPT. This can be done through intensive about how AI-generated content may inadvertently reinforce
socialization of the benefits and how to use ChatGPT among biases, shaping student perspectives and leading to a less
students and teachers. ChatGPT should be promoted within inclusive learning environment.
the school environment through various activities such as While AI offers significant learning benefits, students must
seminars, training, and workshops for students, teachers, and recognize its limitations. Over-reliance on ChatGPT could
parents. The findings from Tlili et al. indicate that although undermine critical thinking if treated as a shortcut rather than
ChatGPT is a valuable tool in education, its usage should be a learning aid. Unlike previous studies, such as Habibi’s,
approached with greater caution, necessitating the which focus on behavioral intentions and infrastructure, this
establishment of more precise guidelines for its safe research emphasizes the need for ethical guidelines that
implementation in educational settings [51]. address AI’s impact on student autonomy and integrity. This
To improve the quality of facilities and access to study thus broadens the literature by advocating for
technology, schools should ensure that technology facilities, responsible and thoughtful use of AI, encouraging students to
such as fast and stable internet access, are available to all view it as a tool that supports rather than replaces their
students. The addition of computer or tablet devices that can learning efforts.
be used to access ChatGPT should also be considered. By considering these implications, students can create a
To improve the design and ease of use of ChatGPT, schools more modern, supportive, and effective learning environment
can provide feedback to the app developers regarding features by utilizing technology such as ChatGPT, thus improving the
that can improve ChatGPT’s ease of use for students. quality of education and students’ readiness to face future
Training and tutorials on using ChatGPT should be offered to challenges.
students and teachers [38].
C. Implications
In developing positive technology use habits, schools must
encourage regular use of ChatGPT in teaching and learning This research contributes to understanding how to
activities, such as daily assignments, research projects, and overcome performance and effort expectations and the role of
extracurricular activities. By doing so, students will get used habit in technology adoption. Despite the low impact of these
to using this technology in their daily academic activities [52]. factors on behavioral intention, future studies should explore
In evaluating and monitoring technology use, schools need ways to enhance students’ perceptions of ChatGPT’s benefits
to conduct regular evaluations of the effectiveness of and ease of use by identifying barriers such as user interface
ChatGPT in the teaching and learning process. Feedback complexity and feature accessibility. Strategies like
from students and teachers should be collected and analyzed simplifying interfaces, offering tutorials, and creating
to improve the use of this technology continuously [53]. The personalized learning experiences can improve usability.
Education Office should oversee and evaluate the programs Additionally, habit’s significant role in influencing intention
implemented in schools. They can create clear policies and and usage suggests the need to cultivate regular, meaningful
guidelines on using technology in education and reward interactions with ChatGPT, potentially through daily
schools and students who demonstrate innovative and academic routines or gamification elements that motivate
effective use of technology. consistent use.

519
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

Teachers’ perspectives on students’ use of ChatGPT vary, Concerns about data privacy, security, and
potential misuse of AI-generated content
with some seeing it as a helpful tool to deepen subject
were highlighted.
understanding. In contrast, others express concern about Addressing Ethical Schools need to establish clear guidelines
students’ over-reliance on technology, fearing it might 3. and Privacy and policies to ensure the ethical use of
weaken critical thinking skills. Teachers also emphasized the Concerns ChatGPT. This includes obtaining parental
consent for students under 18, protecting
importance of clear policies and infrastructure support, student data, and educating students about
including stable internet and proper devices, to ensure the responsible use of AI tools.
effective integration of ChatGPT into the learning process. Access to stable internet connections and
adequate devices was crucial for using
Comprehensive guidance and training for students and ChatGPT effectively.
teachers are essential to optimize ChatGPT’s educational Schools must invest in digital
benefits. 4.
Improving Access infrastructure, including reliable internet
and Infrastructure access and sufficient computer or tablet
This study’s findings suggest that while AI tools like
devices. This ensures that all students have
ChatGPT can enhance learning, ethical considerations must equal opportunities to benefit from AI
guide their use. Schools should implement AI in ways that technologies.
promote critical thinking and responsible use, encouraging
Both students and teachers benefited from
students to explore topics independently rather than relying training sessions on how to use ChatGPT
on AI for quick answers. This approach prevents over- effectively.
reliance on AI, helping students maintain solid analytical and Continuous professional development for
Training and
5. teachers and comprehensive tutorials for
independent thinking skills. Support
students are essential. Schools should
Comprehensive ethical guidelines are essential for the provide ongoing support to help users
responsible use of AI. Schools should ensure data privacy and navigate and utilize ChatGPT’s features
security by using secure platforms and educating students on optimally
Regular monitoring and feedback
data protection practices, such as avoiding sharing sensitive collection were necessary to assess the
information with AI tools. School policies should also include effectiveness of ChatGPT in the learning
guidelines on ethically using AI-generated content, parental process.
Monitoring and Schools should implement a system for
consent for younger students, and monitoring usage to uphold 6.
Evaluation evaluating the impact of ChatGPT on
academic integrity. Additionally, equitable access to AI tools student learning outcomes. Feedback from
is crucial. Partnerships with technology companies and both students and teachers should be used
to make iterative improvements to the
government programs can help under-resourced schools integration of AI tools in education
secure reliable internet and devices. Providing mobile-
friendly or offline-accessible AI tools can further bridge the AI technology, such as ChatGPT, has great potential to
gap for students without consistent internet access, ensuring support student learning. However, students need to
all students can benefit responsibly from AI enhanced understand and apply ethics when using AI to ensure it aligns
learning. with academic standards and personal integrity, as in Fig. 3.
By understanding and applying these implications, First, AI should be seen as a learning aid, not a shortcut to
educators, and policymakers can create a more supportive and completing assignments. Students should use this technology
effective learning environment that leverages the benefits of to deepen their understanding of the material and strengthen
AI technologies like ChatGPT. Table 7 outlines the key their knowledge rather than relying on AI to complete tasks
results and practical recommendations from this study, automatically. This is also related to the principle of
highlighting areas such as enhancing learning efficiency, intellectual honesty, where students must be truthful in using
developing positive technology use habits, and addressing the information generated by AI, including citing the source
ethical concerns. when referencing AI in their work.
Table 7. Result and implications
No. Aspect Result and Implications
This research found that students who
regularly used ChatGPT experienced
quicker and more efficient completion of
assignments and writing tasks.
Enhancing Schools should integrate ChatGPT into
1. Learning daily academic routines, such as homework
Efficiency assignments and project-based learning, to
help students manage their workload more
effectively. This can lead to improved time
management and productivity among
students.
The chronic use of ChatGPT influenced
students’ behavioral intentions and actual
usage positively.
Developing
Schools should incorporate ChatGPT into
Positive
2. various educational activities to encourage
Technology Use
regular and meaningful interactions. This
Habits
could involve gamification elements like
rewards and challenges to motivate
consistent use.
Fig. 3. Ethical guide for students using AI.

520
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

In addition, safeguarding privacy is crucial. Students must technology-enhanced learning. These resources are necessary
avoid entering personal or sensitive information into AI for students to fully benefit from AI tools, impacting their
systems to protect their privacy and that of others. Students overall learning experience and technological readiness.
should also use AI responsibly, ensuring that the content Schools can partner with government programs, tech
produced is not harmful, misleading, or inappropriate. On the companies, and local businesses to address these disparities
other hand, while AI can provide quick answers, students and secure funding for improved infrastructure, including
must continue to rely on their critical thinking skills to better internet and shared devices. Non-profit organizations
analyze information and solve problems. Over-reliance on AI and volunteer groups might also provide free training and
can hinder personal growth in the learning process. technical support to ensure all students can effectively use
If students are unsure about the proper use of AI, they tools like ChatGPT. Implementing mobile friendly or offline
should seek guidance from their teachers. This ensures that accessible versions of AI tools could further support students
AI usage remains in line with school policies and academic who lack internet access at home. Schools can create a more
standards. Moreover, avoiding plagiarism is of the utmost equitable learning environment by focusing on these
importance. Students should create original work and use AI solutions, ensuring all students can benefit from AI in
only as a reference, not as a direct copy without education.
understanding. By applying these ethical principles, students can
AI has its limitations, and students need to be aware of this. maximize the benefits of AI technology while upholding
They should always verify the accuracy of information academic values and personal responsibility. This approach
provided by AI and ensure that the sources used are credible. enhances educational outcomes and prepares students for a
Furthermore, students must maintain digital security by using future where AI plays a significant role in various aspects of
AI properly without compromising the school’s technological life. It’s very possible to become material for future research.
systems. Reporting technical issues or violations is also part As shown in Fig. 4, the implementation of ChatGPT in
of their responsibility. schools follows a structured process that begins with
Lastly, maintaining integrity in the classroom is critical. assessing needs and resources, proceeds through teacher
Students should ensure that AI is only used in class activities, training and curriculum integration, and culminates in
exams, or assessments if permitted by the teacher. This is monitoring and ethical guidelines. In a pilot program at an
important to uphold fairness and integrity in the learning under-resourced rural school, ChatGPT was deployed on
process. shared devices to support English language learning.
Integrating AI tools like ChatGPT in high schools brings Teachers used it to design interactive exercises, while
significant benefits and highlights socio-economic challenges, students accessed it for personalized grammar corrections and
especially in under-resourced areas. Students in these schools vocabulary building. Within three months, students reported
often lack access to reliable internet and devices, which can increased confidence in language skills, and teachers saved
widen the digital divide and limit their exposure to an average of five hours per week on lesson preparation.

Fig. 4. Practical recommendations for implementing ChatGPT.

So, to implement ChatGPT in schools, start by assessing support critical thinking.

needs through surveys on infrastructure and educational The pilot implementation of ChatGPT in an under-
requirements and involve stakeholders like teachers, students, resourced rural school highlighted several challenges that
and parents to address concerns. Provide teacher training informed the development of effective integration strategies.
through workshops and peer learning groups to build skills in Among these, technological limitations posed significant
using ChatGPT for lesson planning and engagement. Begin barriers, such as insufficient access to devices and unreliable
with pilot programs in selected classes, focusing on subjects internet connectivity. To overcome these issues, the school
like language learning or STEM, and create inclusive, organized a structured schedule for shared use of available
student-centred materials. Regularly collect feedback from desktop computers and tablets, ensuring equitable access for
students and teachers to refine the approach. Finally, ethical all students. Additionally, offline resources generated with
guidelines should be established, and students should be ChatGPT were prepared in advance to facilitate continued
educated on using ChatGPT responsibly as a learning tool to learning during connectivity issues. These adaptive measures

521
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

demonstrate how AI tools can be tailored to function This study, which primarily addresses challenges faced by
effectively in low-resource environments. under-resourced schools, presents strategies that can be
Another notable challenge was the reluctance among adapted for urban or better-equipped schools, which typically
teachers to adopt ChatGPT, largely due to a lack of familiarity have greater access to infrastructure, technology, and
with the technology and concerns about its accuracy. To resources. In these environments, ChatGPT can be more
address this, the school organized targeted professional easily integrated into classroom activities with the availability
development workshops to equip educators with the skills of high-speed internet, dedicated devices, and digital learning
and confidence to utilize ChatGPT for lesson planning and platforms. These schools can utilize ChatGPT for
classroom engagement. The formation of peer mentoring personalised learning and to foster collaborative projects,
groups further supported this effort, enabling experienced where students can work together on AI-driven research or
teachers to guide their colleagues through the adoption solve real-world problems in subjects such as STEM,
process. This collaborative approach successfully reduced humanities, or social sciences.
hesitancy and created a more supportive environment for ChatGPT’s ability to provide individualized feedback and
integrating AI technology into teaching practices. support differentiated instruction can be particularly
Engaging students was also a key focus, as initial beneficial in urban schools with larger and more diverse
perceptions of ChatGPT as a formal and complex tool student populations. Teachers can tailor lesson plans to meet
hindered its acceptance. To make the experience more the needs of students with varying proficiency levels within
approachable and enjoyable, the school integrated gamified the same class. For instance, some students might use
elements such as AI-generated quizzes and vocabulary games ChatGPT to deepen their understanding of challenging topics
into the curriculum. A student-led “ambassador” program or receive instant grammar corrections. In contrast, others
was also introduced, where selected students supported their may engage in more creative tasks, such as developing AI-
peers in navigating and utilizing the tool effectively. These driven simulations or producing content-based projects. This
initiatives led to tangible benefits, including increased student flexibility allows ChatGPT to be adapted to a broad spectrum
confidence in language skills and a significant reduction in of teaching approaches and learning goals.
teacher workload, with educators reporting an average of five In capital city schools, where educators may already
hours saved per week on lesson preparation. possess higher technological expertise, professional
For example, Table 8 provides practical prompts for development for ChatGPT integration can be more advanced.
educators during the needs assessment phase. These prompts The focus can shift to using AI for data analysis to track
help ensure comprehensive planning and stakeholder student progress and assess learning outcomes. Rather than
involvement. simply covering the basics, training programs in these schools
could emphasize how AI tools like ChatGPT can enhance
Table 8. Example Prompts for Implementing ChatGPT in Schools
teaching strategies, such as automating the creation of quizzes,
Aspect Example Prompts
“What questions should I include in a survey to generating lesson summaries, and producing advanced
Needs evaluate the technological needs of our school?” instructional materials. Additionally, urban schools could
Assessment “Suggest ways to involve parents and teachers in incorporate ChatGPT into project-based learning, enabling
discussions about integrating ChatGPT in schools.”
“Create a step-by-step guide for training teachers to
students to take more responsibility for their learning,
Training use ChatGPT for lesson planning.” conduct research, collaborate on presentations, and explore
Programs “Design a workshop agenda for introducing ChatGPT new ways to engage with content.
in classroom settings.” Finally, urban schools are in a position to develop more
“Generate a personalized lesson plan using ChatGPT
Curriculum for a STEM class with limited resources.” comprehensive ethical guidelines for AI usage. Since
Integration “How can ChatGPT be used to create language students in these environments will likely encounter AI tools
learning materials for students with diverse needs?” earlier in their academic careers, the guidelines can include
“Develop a feedback form to assess the effectiveness of
teaching digital citizenship, fostering responsible use of AI,
Monitoring & ChatGPT in improving student learning outcomes.”
Evaluation “Suggest a system for monitoring teacher and student and enhancing critical thinking skills. By integrating
engagement with ChatGPT tools.” ChatGPT into digital literacy curricula, these schools can
“Draft guidelines for ensuring academic integrity when promote discussions around the societal impacts of AI, such
Ethical using ChatGPT in schools.”
Guidelines “How can we educate students about responsible use of
as privacy, bias, and misinformation. This proactive approach
ChatGPT in their studies?” not only empowers students to use AI ethically but also
prepares them for the growing role that AI will play in their
These guidelines can guide educators, administrators, and future academic and professional lives.
stakeholders in effectively utilizing ChatGPT to address the
challenges of human resource shortages in schools. With a V. CONCLUSION
structured approach, ChatGPT can be an innovative solution The research highlights that Social Influence, Hedonic
to support teaching and learning activities, reduce teacher Motivation, and Habit significantly impact the intention to
workload, and improve educational accessibility, especially use ChatGPT. Facilitating Conditions, Habit, and Behavioral
in areas with limited infrastructure. Through proper Intention drive actual usage among high school students. To
implementation, such as teacher training, curriculum optimize the benefits of ChatGPT, there needs to be a focus
integration, and the implementation of ethical use policies, on improving its performance expectancy, ease of use, and
ChatGPT not only helps address educational gaps but also supporting facilities. The findings underscore the importance
encourages the creation of more inclusive and adaptive of social dynamics and user satisfaction in adopting
learning experiences for students. educational technologies. Furthermore, numerous challenges

522
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

must be discussed: First, technological infrastructure is learning media innovation in Indonesia,” International Journal of
Multidisciplinary Sciences and Arts, vol. 2, no. 2, pp. 235–241, Dec.
critical that all students have a dependable internet 2023. doi: 10.47709/ijmdsa.v2i2.3333
connection and the required gadgets. Schools must invest in [10] X. An et al., “Modeling English teachers’ behavioral intention to use
digital infrastructure to provide an egalitarian learning artificial intelligence in middle schools,” Educ Inf Technol (Dordr), vol.
28, no. 5, pp. 5187–5208, May 2023. doi: 10.1007/s10639-022-11286-
environment. Second, training and support provide extensive
z
training for instructors and students on how to utilize [11] D. Ayoub, M. Metawie, and M. Fakhry, “AI-ChatGPT usage among
ChatGPT efficiently. This involves creating easy-to-follow users: Factors affecting intentions to use and the moderating effect of
lessons and hosting frequent seminars. Third, privacy and privacy concerns,” MSA-Management Sciences Journal, vol. 0, no. 0,
pp. 120–152, Feb. 2024. doi: 10.21608/msamsj.2024.265212.1054
security, which establish clear data privacy and security [12] C. K. Y. Chan and K. K. W. Lee, “The AI generation gap: Are Gen Z
criteria, are critical for protecting students’ information and students more interested in adopting generative AI such as ChatGPT in
maintaining trust in technology. teaching and learning than their Gen X and millennial generation
teachers?” Smart Learning Environments, vol. 10, no. 1, Dec. 2023. doi:
10.1186/s40561-023-00269-3
CONFLICT OF INTEREST [13] O. Boubker, H. Ben-Saghroune, J. El Bourassi, M. Abdessadek, and R.
Sabbahi, “Examining the impact of OpenAI’s ChatGPT on PhD student
The authors declare no conflict of interest. achievement,” International Journal of Information and Education
Technology, vol. 14, no. 3, pp. 443–451, 2024. doi:
AUTHOR CONTRIBUTIONS 10.18178/ijiet.2024.14.3.2065
[14] C. B. Fontao, M. L. Santos, and A. Lozano, “ChatGPT’s role in the
A. Pratita formulates research objectives and methodology, education system: Insights from the future secondary teachers,”
collects and analyzes data, and composes the research report International Journal of Information and Education Technology, vol.
14, no. 8, pp. 1035–1043, 2024. doi: 10.18178/ijiet.2024.14.8.2131
for publication; T.LM. Suryanto and A. Pratama are academic [15] Y. J. Lee, R. O. Davis, and J. Ryu, “Korean in-service teachers’
advisors. They provide guidance on writing the article by perceptions of implementing Artificial Intelligence (AI) education for
educational standards and ensure the quality of the analysis teaching in schools and their AI teacher training programs,”
International Journal of Information and Education Technology, vol.
and data used. A. Wibowo is the research supervisor, who 14, no. 2, pp. 214–219, 2024. doi: 10.18178/ijiet.2024.14.2.2042
monitors the progress of the research, provides feedback on [16] Venkatesh, Morris, Davis, and Davis, “User Acceptance of information
the analysis and writing, and ensures that the research meets technology: Toward a unified view,” MIS Quarterly, vol. 27, no. 3, p.
academic quality and ethical standards; All authors read and 425, 2003. doi: 10.2307/30036540
[17] V. Venkatesh, J. Y. L. Thong, and X. Xu, “Consumer acceptance and
approved the final manuscript. use of information technology: Extending the unified theory of
acceptance and use of technology,” MIS Quarterly, vol. 36, no. 1, pp.
ACKNOWLEDGMENT 157–178, 2012. doi: doi.org/10.2307/30036540
[18] B. Lund and W. Ting, “Chatting about ChatGPT: How may AI and
We would like to express our deepest gratitude to the GPT impact academia and libraries?” SSRN Electronic Journal, 2023.
University of Pembangunan Nasional “Veteran” Jawa Timur doi: 10.2139/ssrn.4333415
[19] J. Rudolph, S. Tan, and S. Tan, “ChatGPT: Bullshit spewer or the end
for their unwavering support, which enabled us to pursue and of traditional assessments in higher education?” Journal of Applied
complete this research also, appreciation to MSI Laboratory Learning & Teaching, vol. 6, no. 1, Jan. 2023. doi:
UPNVJT for their invaluable assistance and resources. 10.37074/jalt.2023.6.1.9
[20] E. Loh, “ChatGPT and generative AI chatbots: Challenges and
Technical support and dedication have been crucial to opportunities for science, medicine and medical leaders,” BMJ Leader,
achieving this research. p. leader-2023-000797, May 2023. doi: 10.1136/leader-2023-000797
[21] M. Alhwaiti, “Acceptance of artificial intelligence application in the
post-covid era and its impact on faculty members’ occupational well-
REFERENCES being and teaching self efficacy: A path analysis using the UTAUT 2
[1] N. Bhat and A. Dahal, “Self-directed learning, its implementation, and model,” Applied Artificial Intelligence, vol. 37, no. 1, 2023. doi:
challenges: A review,” Nepal Journal of Health Sciences, vol. 3, no. 1, 10.1080/08839514.2023.2175110
pp. 102–115, Dec. 2023. doi: 10.3126/njhs.v3i1.63277 [22] W. Wu, B. Zhang, S. Li, and H. Liu, “Exploring factors of the
[2] S. Schweder and D. Raufelder, “Does changing learning environments willingness to accept AI-assisted learning environments: An empirical
affect student motivation?” Learn Instr, vol. 89, 101829, Feb. 2024. doi: investigation based on the UTAUT model and perceived risk theory,”
10.1016/j.learninstruc.2023.101829 Front Psychol, vol. 13, Jun. 2022. doi: 10.3389/fpsyg.2022.870777
[3] A. J. Head and M. B. Eisenberg, “Truth be told: How college students [23] A. A. A. Zwain, “Technological innovativeness and information
evaluate and use information in the digital age,” SSRN Electronic quality as neoteric predictors of users’ acceptance of learning
Journal, 2010. doi: 10.2139/ssrn.2281485 management system,” Interactive Technology and Smart Education,
[4] R. Dale, “GPT-3: What’s it good for ?” Natural Language Engineering, vol. 16, no. 3, pp. 239–254, Sep. 2019. doi: 10.1108/ITSE-09-2018-
pp. 113–118, 2021. doi: 10.1017/S1351324920000601 0065
[5] R. Zhao, M. M. Yunus, and K. R. M. Rafiq, “The impact of the use of [24] A. S. Jameel, S. N. Abdalla, M. A. Karem, and A. R. Ahmad,
ChatGPT in enhancing students’ engagement and learning outcomes in “Behavioural intention to use e-learning from student’s perspective
higher education: A review,” International Journal of Academic during COVID-19 pandemic,” in Proc. 2020 2nd Annual International
Research in Business and Social Sciences, vol. 13, no. 12, Dec. 2023. Conference on Information and Sciences (AiCIS), IEEE, Nov. 2020, pp.
doi: 10.6007/IJARBSS/v13-i12/20258 165–171. doi: 10.1109/AiCIS51645.2020.00035
[6] K. Hartley, M. Hayak, and U. H. Ko, “Artificial intelligence supporting [25] M. Chen, M. Siu-Yung, C. S. Chai, C. Zheng, and M.-Y. Park, “A pilot
independent student learning: An evaluative case study of ChatGPT study of students’ behavioral intention to use AI for language learning
and learning to code,” Educ Sci (Basel), vol. 14, no. 2, p. 120, Jan. 2024. in higher education,” in Proc. 2021 International Symposium on
doi: 10.3390/educsci14020120 Educational Technology (ISET), IEEE, Aug. 2021, pp. 182–184. doi:
[7] S. Y. Harunasari, “Examining the effectiveness of AI-integrated 10.1109/ISET52350.2021.00045
approach in EFL writing: A case of ChatGPT,” International Journal [26] A. Habibi, M. Muhaimin, B. K. Danibao, Y. G. Wibowo, S. Wahyuni,
of Progressive Sciences and Technologies, vol. 39, no. 2, p. 357, Jul. and A. Octavia, “ChatGPT in higher education learning: Acceptance
2023. doi: 10.52155/ijpsat.v39.2.5516 and use,” Computers and Education: Artificial Intelligence, vol. 5,
[8] F. H. Siregar, B. Hasmayni, and A. H. Lubis, “The analysis of ChatGPT 100190, 2023. doi: 10.1016/j.caeai.2023.100190
usage impact on learning motivation among scout students,” [27] C. S. Chai, P.-Y. Lin, M. S. Jong, Y. Dai, T. K. F. Chiu, and B. Huang,
International Journal of Research and Review, vol. 10, no. 7, pp. 632– “Factors influencing students’ behavioral intention to continue
638, Jul. 2023. doi: 10.52403/ijrr.20230774 artificial intelligence learning,” in Proc. 2020 International Symposium
[9] S. Nita, K. Sussolaikah, and J. D. Aldida, “The role of artificial on Educational Technology (ISET), IEEE, Aug. 2020, pp. 147–150. doi:
intelligence-based technology with ChatGPT as an educational 10.1109/ISET49818.2020.00040

523
 International Journal of Information and Education Technology, Vol. 15, No. 3, 2025

[28] B. N. Obenza et al., “Analyzing university students’ attitude and Technology and Smart Education, Aug. 2023. doi: 10.1108/ITSE-04-
behavior toward AI using the extended unified theory of acceptance 2023-0061
and use of technology model,” American Journal of Applied Statistics [41] S. Hermawan and Amirullah, Metode Penelitian Bisnis Pendekatan
and Economics, vol. 3, no. 1, pp. 99–108, May 2024. doi: Kuantitatif & Kualitatif, 2016.
10.54536/ajase.v3i1.2510 [42] J. F. Hair, C. M. Ringle, and M. Sarstedt, “PLS-SEM: Indeed a silver
[29] Q. AlAjmi, M. A. Al-Sharafi, and A. A. Yassin, “Behavioral intention bullet,” Journal of Marketing Theory and Practice, vol. 19, no. 2, pp.
of students in higher education institutions towards online learning 139–152, Apr. 2011. doi: 10.2753/MTP1069-6679190202
during COVID-19,” 2021, pp. 259–274. doi: 10.1007/978-3-030- [43] S. Yamin and H. Kurniawan, Generasi Baru Mengolah Data Penelitian
67716-9_16 dng Partial Least Square Path Modeling, Salemba Infotek, 2011.
[30] K. Qu and X. Wu, “ChatGPT as a CALL tool in language education: A [44] S. Petter and Y. Hadavi, “With great power comes great responsibility,”
study of hedonic motivation adoption models in English learning ACM SIGMIS Database: the DATABASE for Advances in Information
environments,” Educ Inf Technol (Dordr), Mar. 2024. doi: Systems, vol. 52, no. SI, pp. 10–23, Dec. 2021. doi:
10.1007/s10639-024-12598-y 10.1145/3505639.3505643
[31] A. Strzelecki, “To use or not to use ChatGPT in higher education? A [45] A. Habibi, M. Muhaimin, B. Kopong, Y. Gusti, S. Wahyuni, and A.
study of students’ acceptance and use of technology,” Interactive Octavia, “Computers and education: Artificial intelligence ChatGPT in
Learning Environments, pp. 1–14, May 2023. doi: higher education learning: Acceptance and use,” Computers and
10.1080/10494820.2023.2209881 Education: Artificial Intelligence, vol. 5, no. November, 100190, 2023.
[32] G. García-Murillo, P. Novoa-Hernández, and R. S. Rodríguez, “On the doi: 10.1016/j.caeai.2023.100190
technological acceptance of moodle by higher education faculty—A [46] E. M. Abu-Taieh et al., “Continued intention to use of m-banking in
nationwide study based on UTAUT2,” Behavioral Sciences, vol. 13, Jordan by integrating UTAUT, TPB, TAM and service quality with
no. 5, 2023. doi: 10.3390/bs13050419 ML,” Journal of Open Innovation: Technology, Market, and
[33] S. H. Alshammari and M. H. Alshammari, “Factors affecting the Complexity, vol. 8, no. 3, p. 120, Sep. 2022. doi:
adoption and use of ChatGPT in higher education,” International 10.3390/joitmc8030120
Journal of Information and Communication Technology Education, vol. [47] N. Bozorgkhou, “An internet shopping user adoption model using an
20, no. 1, pp. 1–16, Feb. 2024. doi: 10.4018/IJICTE.339557 integrated TTF and UTAUT: Evidence from Iranian consumers,”
[34] A. A. Hernandez, M. B. Abisado, R. L. Rodriguez, and J. M. R. Management Science Letters, vol. 5, no. 2, pp. 199–204, 2015. doi:
Imperial, “Predicting the use behavior of higher education students on 10.5267/j.msl.2014.12.017
ChatGPT: Evidence from the Philippines,” in Proc. 2023 IEEE [48] R. Wu and J. Lee, “The comparative study on third party mobile
International Conference on Teaching, Assessment and Learning for payment between UTAUT2 and TTF,” Journal of Distribution Science,
Engineering (TALE), IEEE, Nov. 2023, pp. 1–7. doi: vol. 15, no. 11, pp. 5–19, Nov. 2017. doi: 10.15722/jds.15.11.201711.5
10.1109/TALE56641.2023.10398324 [49] N. Ain, K. Kaur, and M. Waheed, “The influence of learning value on
[35] J.-M. Romero-Rodríguez, M.-S. Ramírez-Montoya, M. Buenestado- learning management system use,” Information Development, vol. 32,
Fernández, and F. Lara-Lara, “Use of ChatGPT at university as a tool no. 5, pp. 1306–1321, Nov. 2016. doi: 10.1177/0266666915597546
for complex thinking: Students’ perceived usefulness,” Journal of New [50] C. Morosan and A. DeFranco, “It’s about time: Revisiting UTAUT2 to
Approaches in Educational Research, vol. 12, no. 2, p. 323, Jul. 2023. examine consumers’ intentions to use NFC mobile payments in hotels,”
doi: 10.7821/naer.2023.7.1458 Int J Hosp Manag, vol. 53, pp. 17–29, Feb. 2016. doi:
[36] R. Masa’deh et al., “Antecedents of adoption and usage of ChatGPT 10.1016/j.ijhm.2015.11.003
among Jordanian university students: Empirical study,” International [51] A. Tlili et al., “What if the devil is my guardian angel: ChatGPT as a
Journal of Data and Network Science, vol. 8, no. 2, pp. 1099–1110, case study of using chatbots in education,” Smart Learning
2024. doi: 10.5267/j.ijdns.2023.11.024 Environments, vol. 10, no. 1, p. 15, Feb. 2023. doi: 10.1186/s40561-
[37] C. Yu, J. Yan, and N. Cai, “ChatGPT in higher education: Factors 023-00237-x
influencing ChatGPT user satisfaction and continued use intention,” [52] S. Rice, S. R. Crouse, S. R. Winter, and C. Rice, “The advantages and
Front Educ (Lausanne), vol. 9, May 2024. doi: limitations of using ChatGPT to enhance technological research,”
10.3389/feduc.2024.1354929 Technol Soc, vol. 76, p. 102426, Mar. 2024. doi:
[38] B. Foroughi et al., “Determinants of intention to use ChatGPT for 10.1016/j.techsoc.2023.102426
educational purposes: Findings from PLS-SEM and fsQCA,” Int J Hum [53] J. W. M. Lai, J. Nobile, M. Bower, and Y. Breyer, “Comprehensive
Comput Interact, pp. 1–20, Jun. 2023. doi: evaluation of the use of technology in education—validation with a
10.1080/10447318.2023.2226495 cohort of global open online learners,” Educ Inf Technol (Dordr), vol.
[39] A. Strzelecki, “Students’ acceptance of ChatGPT in higher education: 27, no. 7, pp. 9877–9911, Aug. 2022. doi: 10.1007/s10639-022-10986-
An extended unified theory of acceptance and use of technology,” w
Innov High Educ, vol. 49, no. 2, pp. 223–245, Apr. 2024. doi:
10.1007/s10755-023-09686-1 Copyright © 2025 by the authors. This is an open access article distributed
[40] C. K. Tiwari, Mohd. A. Bhat, S. T. Khan, R. Subramaniam, and M. A. under the Creative Commons Attribution License which permits unrestricted
I. Khan, “What drives students toward ChatGPT? An investigation of use, distribution, and reproduction in any medium, provided the original
the factors influencing adoption and usage of ChatGPT,” Interactive work is properly cited (CC BY 4.0).

524