See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/382946250

ChatGPT usage and attitudes are driven by perceptions of usefulness, ease of

use, risks, and psycho-social impact: a study among university students in the
UAE

Article in Frontiers in Education · August 2024

CITATION READS

1 99

9 authors, including:

Malik Sallam Muhammad Yusuf Al-Shorbagy

University of Jordan Cairo University
183 PUBLICATIONS 7,555 CITATIONS 38 PUBLICATIONS 806 CITATIONS

SEE PROFILE SEE PROFILE

Muna Barakat Sami El Khatib

Applied Science Private University The International University of Beirut
152 PUBLICATIONS 1,197 CITATIONS 100 PUBLICATIONS 373 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Nisreen Alwan on 08 August 2024.

The user has requested enhancement of the downloaded file.

 TYPE Original Research
PUBLISHED 07 August 2024
DOI 10.3389/feduc.2024.1414758

ChatGPT usage and attitudes are

OPEN ACCESS driven by perceptions of
usefulness, ease of use, risks, and
EDITED BY
Hongbiao Yin,
The Chinese University of Hong Kong, China

REVIEWED BY
Xindong Ye,
psycho-social impact: a study
Wenzhou University, China
Ijeoma John-Adubasim,
University of Plymouth, United Kingdom
among university students in
*CORRESPONDENCE
Malik Sallam
the UAE
[email protected]
Diana Malaeb
[email protected]
Malik Sallam 1,2*, Walid Elsayed 3, Muhammad Al-Shorbagy 4,5,
These authors share last authorship
† Muna Barakat 6, Sami El Khatib 7,8, Wissam Ghach 9,
RECEIVED 09 April 2024 Nisreen Alwan 10, Souheil Hallit 11,12,13† and Diana Malaeb 14*†
ACCEPTED 17 July 2024 1
Department of Pathology, Microbiology and Forensic Medicine, School of Medicine, The University
PUBLISHED 07 August 2024
of Jordan, Amman, Jordan, 2 Department of Clinical Laboratories and Forensic Medicine, Jordan
CITATION University Hospital, Amman, Jordan, 3 College of Dentistry, Gulf Medical University, Ajman, United
Sallam M, Elsayed W, Al-Shorbagy M, Arab Emirates, 4 Department of Pharmaceutical Sciences, College of Pharmacy, Gulf Medical
Barakat M, El Khatib S, Ghach W, Alwan N, University, Ajman, United Arab Emirates, 5 Department of Pharmacology and Toxicology, Faculty of
Hallit S and Malaeb D (2024) ChatGPT usage Pharmacy, Cairo University, Cairo, Egypt, 6 Department of Clinical Pharmacy and Therapeutics, Faculty
and attitudes are driven by perceptions of of Pharmacy, Applied Science Private University, Amman, Jordan, 7 Department of Biomedical
usefulness, ease of use, risks, and Sciences, School of Arts and Sciences, Lebanese International University, Bekaa, Lebanon, 8 Center for
psycho-social impact: a study among Applied Mathematics and Bioinformatics (CAMB), Gulf University for Science and Technology (GUST),
university students in the UAE. Hawally, Kuwait, 9 Department of Public Health, Faculty of Communication, Arts and Sciences,
Front. Educ. 9:1414758. Canadian University Dubai, Dubai, United Arab Emirates, 10 College of Health Sciences, Abu Dhabi
doi: 10.3389/feduc.2024.1414758 University, Abu Dhabi, United Arab Emirates, 11 School of Medicine and Medical Sciences, Holy Spirit
University of Kaslik, Jounieh, Lebanon, 12 Department of Psychology, College of Humanities, Effat
COPYRIGHT
University, Jeddah, Saudi Arabia, 13 Applied Science Research Center, Applied Science Private
© 2024 Sallam, Elsayed, Al-Shorbagy, Barakat, University, Amman, Jordan, 14 Department of Pharmacy Practice, College of Pharmacy, Gulf Medical
El Khatib, Ghach, Alwan, Hallit and Malaeb. University, Ajman, United Arab Emirates
This is an open-access article distributed
under the terms of the Creative Commons
Attribution License (CC BY). The use,
distribution or reproduction in other forums is
permitted, provided the original author(s) and Background: The use of ChatGPT among university students has gained a
the copyright owner(s) are credited and that
the original publication in this journal is cited,
recent popularity. The current study aimed to assess the factors driving the
in accordance with accepted academic attitude and usage of ChatGPT as an example of generative artificial intelligence
practice. No use, distribution or reproduction (genAI) among university students in the United Arab Emirates (UAE).
is permitted which does not comply with
these terms. Methods: This cross-sectional study was based on a previously validated
Technology Acceptance Model (TAM)-based survey instrument termed
TAME-ChatGPT. The self-administered e-survey was distributed by emails for
students enrolled in UAE universities during September–December 2023 using
a convenience-based approach. Assessment of the demographic and academic
variables, and the TAME-ChatGPT constructs’ roles in ChatGPT attitude and
usage was conducted using univariate followed by multivariate analyses.
Results: The final study sample comprised 608 participants, 91.0% of whom
heard of ChatGPT while 85.4% used ChatGPT before the study. Univariate
analysis indicated that a positive attitude to ChatGPT was associated with the
three TAME-ChatGPT attitude constructs namely, lower perceived risks, lower
anxiety, and higher scores on the attitude to technology/social influence.
For the ChatGPT usage, univariate analysis indicated that positive attitude to
ChatGPT use was associated with being male, Arab in nationality, and lower
point grade average (GPA) as well as the four ChatGPT usage constructs namely,
higher perceived usefulness, lower perceived risks of use, higher scores on the
behavior/cognitive construct and higher scores on the ease-of-use construct.

Frontiers in Education 01 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

In multivariate analysis, only TAME-ChatGPT constructs explained the variance

in attitude towards ChatGPT (80.8%) and its usage (76.9%).
Conclusion: The findings indicated that ChatGPT usage is commonplace among
university students in the UAE. The determinants of use included the perceived
usefulness, lower perceived risks, cognitive and behavioral factors, perceived
ease of use, while the attitude was determined by lower perceived risks, lower
anxiety, and higher scores for attitude to technology/social influence. These
factors should be considered for understanding the motivators for successful
adoption of genAI including ChatGPT in higher education.

KEYWORDS

AI in education, higher education, large language models, attitude, ChatGPT

1 Introduction the needed support due to perceived barriers or misconceptions

(Karen et al., 2023; Ng et al., 2023; Barakat et al., 2024). Fourth, the
The integration of technology is becoming an indispensable rapid emergence and evolution of genAI models could surpass the
component to improve the quality of higher education (Haleem et al., pace of developing policies and regulations for successful
2022; Criollo-C et al., 2023; Okoye et al., 2023). Recently, the implementation and responsible use of these tools (Chan, 2023;
incorporation of various generative artificial intelligence (genAI) Dempere et al., 2023; Lim et al., 2023). In turn, this could potentially
models in education received a significant attention (Kamalov et al., create significant challenges in establishing standardized practices for
2023; King and Prasetyo, 2023; Mijwil et al., 2023; Yu and Guo, 2023). governing higher education. Fifth, the impact of genAI on the job
The genAI role in higher education represents a paradigm shift which market necessitates a thorough re-evaluation of the competencies and
could redefine the fundamental aspects of teaching and learning skills acquired during higher education to ensure the preparation of
methodologies (Ouyang and Jiao, 2021; Yu, 2024). graduates capable to adapt in a rapidly changing work environment
The emergence of genAI exemplified by popular tools such as (Bukartaite and Hooper, 2023; Gupta, 2024; Tayan et al., 2024).
ChatGPT (OpenAI, San Francisco, CA), could mark a revolution Educators and students have been shown to increasingly utilize
rather than an evolution which could reshape the entire educational genAI models with ChatGPT being among the most popular of these
landscape (Caleb et al., 2023; Fütterer et al., 2023; Johnson, 2023). The tools (Ansari et al., 2023; Ibrahim et al., 2023; von Garrel and Mayer,
potential educational benefits of genAI including ChatGPT especially 2023; Abdaljaleel et al., 2024). Therefore, it is important to understand
in health education attracted significant research attention within a how these genAI models are perceived and utilized especially among
short time span (Ogunleye et al., 2024; Sallam, 2024; Sallam et al., students. Such an inquiry could be viewed as a critical factor for
2024). Generative AI tools are characterized by a remarkable ability to successful implementation of genAI models including ChatGPT into
understand and respond to natural language queries (Bandi et al., the educational framework (Chan and Hu, 2023; Sallam et al., 2024).
2023). On one hand, these capabilities of genAI models offer The implications of this research area are far-reaching.
innovative educational benefits. These benefits include enhancing Understanding the factors driving the genAI adoption in higher
personalized learning experiences and providing realistic simulations education can inform the development of effective implementation
which would help to create an engaging educational content; thus, strategies (Kamalov et al., 2023). Additionally, this area of research
improving student engagement and learning outcomes (Kurtz et al., could shed light on the broader implications of genAI for the future
2024; Salinas-Navarro et al., 2024a,b). For example, (Kıyak and of higher education and the job market (Michel-Villarreal et al., 2023).
Emekli, 2024) showed the efficiency of ChatGPT in generating Such a quest involves the assessment of demographic, academic,
medical multiple-choice questions (MCQs) in a recent review. psychological, social, and economic aspects driving the attitude
However, the same tools, including ChatGPT, pose valid challenges towards this novel technology (Farina and Lavazza, 2023; Ibrahim
and ethical concerns igniting controversy in aspects such as bias, et al., 2023; Zarifhonarvar, 2023; Abdaljaleel et al., 2024). In addition,
cybersecurity, plagiarism and academic dishonesty (Michel-Villarreal this investigation can help to embrace genAI tools as constructive
et al., 2023; Sallam, 2023; Salazar et al., 2024; Williams, 2024). assets within educational settings, rather than viewing this inevitable
Specifically, the concerns regarding genAI include but are not technology as a challenge (de Winter et al., 2023).
limited to the following aspects. First, a decline in the critical thinking A comprehensive framework for assessing the determinants of
and problem-solving skills can occur among students due to the over- adopting a novel technology is the Technology Acceptance Model
reliance on technology (Sallam et al., 2023b). Second, variability in the (TAM) (Davis, 1989; Bagozzi et al., 1992; Marangunić and Granić, 2015).
ability to get access to novel technologies within and between different Based on the TAM framework, a recently developed and validated
societies can put students lacking such an ability at a disadvantage instrument termed “TAME-ChatGPT” described several factors as
creating a digital divide (Ragnedda and Muschert, 2013; Kitsara, drivers of the attitude to ChatGPT and its usage among university
2022). Third, genAI integration into educational practices requires students (Sallam et al., 2023a). These factors include the perceived
adaptation from the educators, who may display hesitancy or lack of usefulness, behavioral and cognitive factors, general perceived risks and

Frontiers in Education 02 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

the perceived risks of use, the perceived ease of use, anxiety, attitude to proportion or apparent prevalence with specified precision available
technology and social influence (Sallam et al., 2023a). from Epitools – Epidemiological Calculators (2024).
Based on the TAME-ChatGPT tool, the current study aimed to
unravel the factors driving the adoption of ChatGPT among university
students in the United Arab Emirates (UAE). The UAE presents a 2.2 Ethical considerations
unique setting for this investigation, given its diverse cultural
composition and rapidly evolving higher education landscape with This study was approved by the Institutional Review Board (IRB)
aspiration to achieve top tier quality in education (Badry, 2019). The at Gulf Medical University (Reference number: IRB-COD-FAC-49-
UAE has placed AI at the core of its national agenda, establishing a APRIL-2023). Obtaining the informed consent to participation was
deep commitment to embedding AI within its economic and ensured by the inclusion of a mandatory item at the beginning of the
technological strategies (Alkhaldi and Altaei, 2021; Shwedeh et al., electronic survey to explicitly indicate consent for participation.
2024). A pioneering step in this direction was the early establishment
of the Ministry of State for Artificial Intelligence, Digital Economy &
Remote Work Applications. As the first entity of its kind worldwide, 2.3 Survey instrument
this ministry coordinates AI governance and policy, positioning the
UAE as a leader in the global AI landscape (Dahabreh, 2023). The electronic survey started with an introductory section which
Furthermore, the UAE hosts the world’s first dedicated AI research outlined the study objectives. This was followed by the mandatory
university at the graduate level, highlighting its strategic educational informed consent item: “Do you agree to participate in this study?”
initiatives aimed at developing a proficient workforce to meet the Agreement with “yes” as an answer allowed the participant to proceed
demands of an AI-driven future (Science/AAAS Custom Publishing to the subsequent survey sections, whereas disagreement as indicated
Office, 2023). The current study implications could help in contributing by “no” response resulted in closure of the survey.
to the growing literature assessing the determinants of generative AI The following section assessed the socio-demographic and academic
implementation in higher education. Additionally, this study sought to data including the following variables: (1) age (as a scale variable); (2) sex
provide insights that can guide educators and academic policymakers (male vs. female); (3) nationality (Arab vs. non-Arab); (4) college/faculty
regarding the students’ perspectives on ChatGPT which can affiliation (health-related (Health Sciences and Public Health colleges), vs.
consequently help to enrich their educational experience. non-health-related (Art and Sciences, Law, Business, Engineering,
Military, Electrical Engineering, Communication, Arts, and Sciences, and
Social Sciences colleges)); (5) current educational level (undergraduate vs.
2 Methods postgraduate); and (6) the latest self-reported GPA (optional item), later
classified into four categories as follows: <2.50, 2.50–2.99, 3.00–3.49, and
2.1 Study design 3.50–4.00.
The next section comprised two preliminary questions: first, “Have
This study adopted a validated survey instrument based on the you heard of ChatGPT before the study?” (Yes vs. No), where a “No”
technology acceptance model (TAM) and specifically tailored to response led to the survey submission. A “Yes” response led to the second
measure the attitude towards ChatGPT among university students question, “Have you used ChatGPT before the study?” (Yes vs. No).
(Sallam et al., 2023a). The survey instrument validity was confirmed Respondents who had not used ChatGPT were directed to a set of 13
in a recently published multinational study among university students attitude TAME-ChatGPT scale questions, whereas those who had used
in five Arab countries (Abdaljaleel et al., 2024). ChatGPT proceeded to a comprehensive set of 25 TAME-ChatGPT items
This study utilized a self-administered electronic survey, addressing both attitude to ChatGPT and its usage.
distributed via email to university students in the UAE. The survey The survey items comprising the constructs of the TAME-ChatGPT,
employed a non-probability, convenience sampling approach, hosted are outlined in Appendix. Each item was evaluated using a 5-point Likert
using Google Forms, and distributed by the authors based in the UAE scale, where “strongly agree” was scored as 5, “agree” as 4, “neutral/no
(W.E., M.A.-S., W.G., N.A., and D.M.). The questionnaire was offered opinion” as 3, “disagree” as 2, and “strongly disagree” as 1. For the TAME-
simultaneously in both Arabic and English languages to accommodate ChatGPT items indicative of a negative attitude (perceived risk, anxiety,
the linguistic preferences and cultural diversity among university and perceived risk of use), the scoring was reversed.
students in the UAE. The survey was accessible from 20 September The attitude scale encompassed three constructs: a perceived risk
2023 to 8 December 2023. Participation in the study was entirely sub-scale with 5 items, an anxiety sub-scale with 3 items, and an
voluntary, with no incentives for participation. To reduce the effect of attitude to technology/social influence sub-scale with 5 items. The
non-response bias, responding to all items were mandatory for usage scale comprised four constructs: perceived usefulness sub-scale
successful completion of the questionnaire with the exception of self- with 6 items, behavior/cognitive factors sub-scale with 3 items,
reported latest cumulative grade point average (GPA). perceived risk of ChatGPT use sub-scale with 3 items, two of which
The minimum required sample size was calculated at 384 based on were also present in the perceived risk construct, and perceived ease
the formula: n = (Z2 × P × (1 − P))/e2, where: Z = 1.96 for 95% confidence of use subs-scale with 2 items (Appendix).
interval (CI), P as the expected true proportion (set at 50%), and e as the
desired precision (set at ±0.05), and the latest estimate of the total
number of university students in the UAE as retrieved from the UAE 2.4 Statistical and data analyses
Ministry of Education official website in the academic year 2019/2020
(The UAE Ministry of Education, 2024). Calculation of the minimum Statistical analyses were conducted using IBM SPSS Statistics
sample size was done using the EPITOOLS sample size to estimate a Version 26.0 (Armonk, NY: IBM Corp). The association between

Frontiers in Education 03 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

categorical variables was evaluated using the Chi-squared test (χ2). 3 Results
The Chi-squared test was selected for its effectiveness in
determining statistical significance between categorical variables 3.1 General features of the study sample
in a contingency table. This choice was supported by the sufficient
sample size, which ensured that the expected frequency in each A total of 608 responses were collected over the period 20
cell of the table was adequate. For the analysis involving September 2023 to 8 December 2023. Most of the participants were
categorical and scale variables, the Mann–Whitney U (M-W) and females, less than 21 years in age, Arabs, enrolled in non-health-
Kruskal-Wallis H (K-W) tests were employed. These related colleges, and undergraduates. For age as a scale variable, the
non-parametric tests are appropriate for datasets where a normal overall mean age of the participants was 20.9 ± 3.5 (median: 20,
distribution cannot be assumed. Specifically, the M-W test was interquartile range (IQR): 19–22). The latest self-reported cumulative
utilized to compare two independent groups when the dependent GPA data were available from 520 participants out of the 608
variable was a scale variable that is not normally distributed. For participants (85.5%). The vast majority of participants heard of
comparisons involving more than two groups, the K-W test, ChatGPT (91.0%) or used ChatGPT before the study (85.4%, Table 1).
which extends the M-W test, was applied. The selection of these Subsequent analysis was conducted among those who heard of
tests was done following the determination of non-normal ChatGPT for the attitude constructs (n = 553), and among those who
distribution of the scale variables via the Kolmogorov–Smirnov indicated ChatGPT usage before the study for the usage constructs
test (p < 0.001 for all). The level of statistical significance was (n = 472).
determined at p < 0.050.
For the multivariate analysis, predictor variables were included
based on p < 0.100 in univariate analysis. The selection of a less 3.2 Analysis of TAME-ChatGPT usage
stringent threshold allowed an exploratory approach in model constructs
building for the identification of potentially important variables
which could have been overlooked with a more conservative p The highest average score for the TAME-ChatGPT usage
value cutoff. constructs was observed for the ease of use construct with a mean
Multivariate regression analysis was employed to assess the score of 4.36 ± 0.74 followed by the perceived usefulness construct
influence of multiple predictors simultaneously, accounting for their with a mean score of 3.97 ± 0.80, behavior/cognitive construct with a
interdependencies. The overall significance of the regression model mean score of 3.73 ± 0.97, and finally the perceived risk of use
was evaluated, which was crucial to determine whether the set of construct with a mean score of 2.06 ± 0.77 (Figure 1).
variables in the model significantly predicted the outcome variable, Univariate analysis of the demographic factors associated with
compared to a model with no predictors. This evaluation was reported each TAME-ChatGPT usage construct revealed statistically
as an F-test in the Analysis of Variance (ANOVA) table provided in significant higher scores among males, Arabs, participants in
regression analysis output, testing the null hypothesis that no non-health-related colleges, and postgraduates that for both the
relationship existed between the dependent and independent perceived usefulness construct and the behavior/cognitive construct
variables. To ensure the reliability of the regression analysis, the (Table 2).
Variance Inflation Factor (VIF) was used to assess multicollinearity
among predictors. A VIF value >3.0 was used as a conservative
threshold to flag variables that might excessively inflate variances, 3.3 Attitude towards ChatGPT based on
which helped to prevent the inclusion of highly correlated variables TAME-ChatGPT constructs
that could distort the true relationship between predictors and
the outcome. The highest average score for the TAME-ChatGPT attitude
Scores for each construct of the TAME-ChatGPT scale were sub-scales was observed for the technology/social influence
calculated by dividing the total scores by the number of items within construct with a mean score of 3.95 ± 0.82 followed by the
that construct, resulting in a score range of 1–5. The overall TAME- perceived risk construct with a mean score of 2.08 ± 0.79, and
ChatGPT scores were based on the mean of the scores for scale items finally the anxiety construct with a mean score of 2.07 ± 0.92
divided by the number of items in each scale. The scoring (Figure 2).
classification for both the TAME-ChatGPT and its individual Univariate analysis of the demographic factors associated with
constructs was categorized as follows: a score range of 1.00 to 2.33 each TAME-ChatGPT attitude constructs revealed statistically
indicated disagreement (negative), 2.34 to 3.67 indicated neutral significant higher scores among males, Arabs, and participants in
position, and 3.68 to 5.00 indicated agreement (positive). The internal non-Health-related colleges for attitude to technology/social influence
consistency of the seven TAME-ChatGPT constructs were ensured construct (Table 3).
by the following Cronbach’s α values: perceived usefulness = 0.888,
behavior/cognitive factors = 0.796, perceived risk of use = 0.638,
perceived ease of use = 0.779, perceived risk = 0.846, anxiety = 0.867, 3.4 Univariate analysis of the attitude and
and attitude to technology/social influence = 0.904. The Cronbach’s α usage of ChatGPT based on
value for the overall TAME-ChatGPT usage scale was 0.797, while TAME-ChatGPT constructs
the value for the overall attitude scale was 0.736. The calculated
Cronbach’s α values indicated an acceptable level of consistency To assess the overall factors influencing the usage of ChatGPT,
within the TAME-ChatGPT constructs (Tavakol and Dennick, 2011). univariate analysis revealed that the following demographic variables

Frontiers in Education 04 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

TABLE 1 General features of the study sample (N = 608).

Variable Category Count Percentage

Age ≤ 20 years 347 57.1

> 20 years 261 42.9

Sex Male 287 47.2

Female 321 52.8

Nationality Arab 412 67.8

Non-Arab 196 32.2

College a
Health-related 155 25.5

Non-Health-related 453 74.5

Latest self-reported cumulative GPAb < 2.50 54 10.4

2.50–2.99 90 17.3

3.00–3.49 173 33.3

3.50–4.00 203 39.0

Educational level Undergraduate 580 95.4

Postgraduate 28 4.6

Have you heard of ChatGPT before this study? Yes 553 91.0

No 55 9.0

Have you used ChatGPT before this study? Yes 472 85.4

No 81 14.6
a
College: Health-related included Health Sciences and Public Health colleges, while non-Health-related included Art and Sciences, Law, Business, Engineering, Military, Electrical Engineering,
Communication, Arts, and Sciences, and Social Sciences colleges.
b
GPA: Grade point average with information available from 520 participants.

FIGURE 1
Box plots showing the distribution of scores across the four TAME-ChatGPT usage constructs.

were significantly associated with higher ChatGPT usage scores: being a associated with higher ChatGPT usage scores as follows: higher
male, an Arab in nationality, and lower self-reported latest GPA (Table 4). perceived usefulness, higher scores of the behavior/cognitive factors,
Additionally, the four TAME-ChatGPT constructs were significantly lower perceived risk of use, and higher perceived ease of use (Table 4).

Frontiers in Education 05 frontiersin.org

Frontiers in Education

Sallam et al.
TABLE 2 The demographic determinants of ChatGPT use based on TAME-ChatGPT usage constructs.

Variable Category Average Average Average Average

perceived behavior/ perceived risk of perceived ease
usefulness sub- cognitive factors use sub-scale of use sub-scale
scale sub-scale
Mean ± SDc p valued Mean ± SD p value Mean ± SD p value Mean ± SD p value
Age ≤ 20 years 3.95 ± 0.78 0.292 3.72 ± 0.97 0.730 2.00 ± 0.73 0.076 4.35 ± 0.71 0.361

> 20 years 4.01 ± 0.83 3.74 ± 0.97 2.15 ± 0.82 4.39 ± 0.77

Sex Male 4.14 ± 0.77 <0.001 3.89 ± 0.96 <0.001 2.04 ± 0.82 0.569 4.39 ± 0.75 0.329

Female 3.82 ± 0.80 3.57 ± 0.96 2.07 ± 0.72 4.34 ± 0.73

Nationality Arab 4.05 ± 0.81 <0.001 3.86 ± 0.92 <0.001 2.08 ± 0.82 0.692 4.38 ± 0.76 0.120

Non-Arab 3.81 ± 0.76 3.45 ± 1.02 2.02 ± 0.65 4.32 ± 0.68

06

Collegea Health-related 3.81 ± 0.81 0.013 3.44 ± 1.03 <0.001 2.11 ± 0.79 0.428 4.28 ± 0.78 0.131

Non-Health-related 4.02 ± 0.79 3.82 ± 0.94 2.04 ± 0.77 4.39 ± 0.72

Cumulative GPA b
< 2.50 4.23 ± 0.79 <0.001 4.17 ± 0.83 <0.001 1.91 ± 0.76 0.352 4.33 ± 0.87 0.854

2.50–2.99 4.11 ± 0.78 3.85 ± 0.92 2.12 ± 0.80 4.46 ± 0.62

3.00–3.49 4.06 ± 0.82 3.85 ± 1.01 1.99 ± 0.76 4.35 ± 0.75

3.50–4.00 3.79 ± 0.81 3.56 ± 0.93 2.11 ± 0.77 4.38 ± 0.71

Educational level Undergraduate 3.97 ± 0.79 0.013 3.72 ± 0.97 0.001 2.06 ± 0.76 0.428 4.38 ± 0.72 0.131

Postgraduate 3.99 ± 1.01 3.83 ± 1.11 2.10 ± 1.01 4.07 ± 0.98

a
College: Health-related included Health Sciences and Public Health colleges, while non-Health-related included Art and Sciences, Law, Business, Engineering, Military, Electrical Engineering, Communication, Arts, and Sciences, and Social Sciences colleges.
b
GPA: Grade point average.
c
SD: Standard deviation.
d
p value: Calculated using the Mann–Whitney and Kruskal Wallis tests.

10.3389/feduc.2024.1414758
Statistically significant p values are highlighted in bold style.
frontiersin.org
Sallam et al. 10.3389/feduc.2024.1414758

FIGURE 2
Box plots showing the distribution of scores across the three TAME-ChatGPT attitude constructs.

TABLE 3 The demographic determinants of attitude to ChatGPT based on TAME-ChatGPT attitude constructs.

Variable Category Average Average Average attitude

perceived risk anxiety sub- to technology/
sub-scale scale social influence
sub-scale
Mean ± SDc p valued Mean ± SD p value Mean ± SD p value
Age ≤ 20 years 2.02 ± 0.76 0.106 2.01 ± 0.90 0.058 3.92 ± 0.80 0.175

> 20 years 2.15 ± 0.82 2.16 ± 0.94 3.99 ± 0.85

Sex Male 2.07 ± 0.81 0.673 2.13 ± 0.96 0.199 4.12 ± 0.77 <0.001

Female 2.08 ± 0.76 2.01 ± 0.89 3.79 ± 0.83

Nationality Arab 2.07 ± 0.83 0.500 2.06 ± 0.94 0.770 4.02 ± 0.81 0.001

Non-Arab 2.08 ± 0.69 2.07 ± 0.89 3.79 ± 0.83

Collegea Health-related 2.08 ± 0.82 0.969 2.09 ± 0.97 0.988 3.77 ± 0.84 0.001

Non-Health-related 2.07 ± 0.78 2.06 ± 0.90 4.01 ± 0.81

Cumulative GPAb < 2.50 1.88 ± 0.75 0.270 2.01 ± 0.90 0.457 4.22 ± 0.78 0.022

2.50–2.99 2.18 ± 0.84 2.23 ± 0.98 3.92 ± 0.87

3.00–3.49 2.04 ± 0.79 2.01 ± 0.89 4.03 ± 0.84

3.50–4.00 2.07 ± 0.76 2.07 ± 0.93 3.86 ± 0.83

Educational level Undergraduate 2.08 ± 0.78 0.139 2.08 ± 0.92 0.138 3.94 ± 0.81 0.253

Postgraduate 1.92 ± 0.98 1.87 ± 1.00 3.98 ± 1.10

a
College: Health-related included Health Sciences and Public Health colleges, while non-Health-related included Art and Sciences, Law, Business, Engineering, Military, Electrical Engineering,
Communication, Arts, and Sciences, and Social Sciences colleges.
b
GPA: Grade point average.
c
SD: Standard deviation.
d
p value: Calculated using the Mann–Whitney and Kruskal Wallis tests.
Statistically significant p values are highlighted in bold style.

For the attitude towards ChatGPT, univariate analysis showed that score with lower perceived ChatGPT risk, lower anxiety, and higher
the three TAME-ChatGPT constructs were the only variables having scores on attitude to technology/social influence being linked with
significant associations with the overall TAME-ChatGPT attitude higher attitude scores (Table 5).

Frontiers in Education 07 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

TABLE 4 Univariate analysis associated with TAME-ChatGPT usage.

Variable Category Overall TAME-ChatGPT usage categoriesb

Negative Neutral Positive p value, χ2
Count (%) Count (%) Count (%)
Age ≤ 20 years 1 (0.3) 179 (62.2) 108 (37.5) 0.456, 1.571

> 20 years 2 (1.1) 107 (58.2) 75 (40.8)

Sex Male 2 (0.9) 123 (53.2) 106 (45.9) 0.006, 10.316

Female 1 (0.4) 163 (67.6) 77 (32.0)

Nationality Arab 3 (0.9) 177 (54.6) 144 (44.4) <0.001, 16.013

Non-Arab 0 (0) 109 (73.6) 39 (26.4)

College Health-related 0 (0) 76 (67.9) 36 (32.1) 0.144, 3.876

Non-Health-related 3 (0.8) 210 (58.3) 147 (40.8)

Cumulative GPA < 2.50 1 (2.4) 20 (47.6) 21 (50.0) 0.049, 12.658

2.50–2.99 0 (0) 38 (53.5) 33 (46.5)

3.00–3.49 1 (0.7) 71 (53.0) 62 (46.3)

3.50–4.00 1 (0.6) 113 (68.1) 52 (31.3)

Educational level Undergraduate 2 (0.4) 271 (60.4) 176 (39.2) 0.056, 5.761

Postgraduate 1 (4.3) 15 (65.2) 7 (30.4)

Perceived usefulness categories Disagreement 3 (18.8) 13 (81.3) 0 (0) <0.001, 201.836

Neutral 0 (0) 126 (97.7) 3 (2.3)

Agreement 0 (0) 147 (45.0) 180 (55)

Behavior/cognitive factors categories Disagreement 3 (5.4) 53 (94.6) 0 (0) <0.001, 132.198

Neutral 0 (0) 103 (88.8) 13 (11.2)

Agreement 0 (0) 130 (43.3) 170 (56.7)

Perceived risk of use categories a

Disagreement 1 (0.3) 214 (62.4) 128 (37.3) <0.001, 78.008

Neutral 0 (0) 69 (57.0) 52 (43.0)

Agreement 2 (25.0) 3 (37.5) 3 (37.5)

Perceived ease of use categories Disagreement 2 (28.6) 5 (71.4) 0 (0) <0.001, 124.765

Neutral 0 (0) 60 (93.8) 4 (6.3)

Agreement 1 (0.2) 221 (55.1) 179 (44.6)

a
Agreement indicated lower perceived risk of use based on reverse coding of the items.
b
TAME-ChatGPT usage categories based on the average scores were classified as 1.00–2.33 (negative), 2.34–3.67 (neutral), and 3.68–5.00 (positive).
Statistically significant p values are highlighted in bold style.

3.5 Multivariate analysis of the 3.6 Multivariate analysis of the

determinants of ChatGPT usage determinants of attitude towards ChatGPT

The regression model for the predictors of ChatGPT usage For the predictors of the attitude towards ChatGPT, the regression
showed a high degree of explanatory power with an R2 value of model showed a high degree of explanatory power with an R2 value of
0.774 indicating that 77.4% of the variation in the usage of 0.808. Using this model, the three TAME-ChatGPT attitude constructs
ChatGPT were accounted for by the predictors included in the were associated with the attitude scores as follows: lower perceived
model. Nationality was the only demographic variable with a risk (B = 0.418, p < 0.001), lower anxiety scores (B = 0.479, p < 0.001),
significant association with TAME-ChatGPT usage score and higher scores in the attitude to technology/social influence
(p = 0.025), suggesting that being Arab in nationality was linked construct (B = 0.413, p < 0.001, Table 7).
with higher usage scores.
On the other hand, the four TAME-ChatGPT usage constructs
were associated with the overall TAME-ChatGPT usage scores as 4 Discussion
follows: higher perceived usefulness (B = 0.398, p < 0.001), higher
scores on the behavior/cognitive factors (B = 0.276, p < 0.001), lower The current study highlighted a notable increase in the usage of
perceived risk of use (B = 0.265, p < 0.001), and higher perceived ease ChatGPT among university students, with a substantial increase
of use scores (B = 0.368, p < 0.001, Table 6). compared to earlier studies from different world regions. The findings

Frontiers in Education 08 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

TABLE 5 Univariate analysis associated with the overall TAME-ChatGPT attitude score.

Variable Category Overall TAME-ChatGPT attitude categoriesc

Negative Neutral Positive p value, χ2
Count (%) Count (%) Count (%)
Age ≤ 20 years 1 (0.3) 179 (62.2) 108 (37.5) 0.086, 4.914

> 20 years 2 (1.1) 107 (58.2) 75 (40.8)

Sex Male 2 (0.9) 123 (53.2) 106 (45.9) 0.126, 4.143

Female 1 (0.4) 163 (67.6) 77 (32.0)

Nationality Arab 3 (0.9) 177 (54.6) 144 (44.4) 0.787, 0.479

Non-Arab 0 (0) 109 (73.6) 39 (26.4)

College Health-related 0 (0) 76 (67.9) 36 (32.1) 0.067, 5.410

Non-Health-related 3 (0.8) 210 (58.3) 147 (40.8)

Cumulative GPA < 2.50 1 (2.4) 20 (47.6) 21 (50.0) 0.184, 8.814

2.50–2.99 0 (0) 38 (53.5) 33 (46.5)

3.00–3.49 1 (0.7) 71 (53.0) 62 (46.3)

3.50–4.00 1 (0.6) 113 (68.1) 52 (31.3)

Educational level Undergraduate 2 (0.4) 271 (60.4) 176 (39.2) 0.437, 1.655

Postgraduate 1 (4.3) 15 (65.2) 7 (30.4)

Perceived risk categoriesa Disagreement 148 (41.5) 209 (58.5) 0 (0) <0.001, 201.097

Neutral 7 (3.9) 159 (89.3) 12 (6.7)

Agreement 0 (0) 9 (50.0) 9 (50)

Anxiety categoriesb Disagreement 152 (39.1) 237 (60.9) 0 (0) <0.001, 227.671

Neutral 3 (2.3) 123 (92.5) 7 (5.3)

Agreement 0 (0) 17 (54.8) 14 (45.2)

Attitude to technology/social Disagreement 13 (65.0) 7 (35.0) 0 (0)

influence categories

Neutral 56 (35.0) 104 (65.0) 0 (0) <0.001, 29.956

Agreement 86 (23.1) 266 (71.3) 21 (5.6)

a
Agreement indicated lower perceived risk based on reverse coding of the items.
b
Agreement indicated lower anxiety based on reverse coding of the items.
c
TAME-ChatGPT usage categories based on the average scores were classified as 1.00–2.33 (negative), 2.34–3.67 (neutral), and 3.68–5.00 (positive).
Statistically significant p values are highlighted in bold style.

of this study revealed that 85% of the participating university students usage among 19% of the participants (Gulati et al., 2024). Taken
in the UAE have already used ChatGPT. An early study conducted together, these results highlight a noticeable rise in the adoption of
during February–March 2023 among university students in health ChatGPT among university students and its evolving status to become
schools in Jordan reported that only 11% used ChatGPT at the time a normal practice within this demographic group.
(Sallam et al., 2023a). A subsequent multinational study that involved In this study, the univariate analysis identified a positive
university students in five Arab countries (Iraq, Kuwait, Egypt, correlation between the attitude to technology/social influence in the
Lebanon, and Jordan) during April–August 2023 reported ChatGPT context of general attitude to ChatGPT and the following demographic
usage at a rate of 25% (Abdaljaleel et al., 2024). groups: being an Arab student, being a male participant, and affiliation
In a culturally different context, a study that was conducted during in non-health colleges. This association might indicate the interplay
June–July 2023 revealed that 39% of medical students across Germany, of cultural and demographic factors in the adoption of new
Austria, and Switzerland previously engaged with AI-based chatbots technologies such as ChatGPT. For example, Arab students might
including ChatGPT (Weidener and Fischer, 2024). Another exhibit more positive attitude to innovative AI technologies possibly
multinational study among academics and university students in due to its utility in overcoming language barriers (Chen, 2023; Mijwil
Brazil, India, Japan, United Kingdom, and United States, which was et al., 2023b; Barwise et al., 2024). Nevertheless, this justification
conducted in January 2023 indicated that a majority of students intend remains tentative considering the studies showing inferior
to use ChatGPT for support in university assignments and anticipate performance of ChatGPT in non-English languages (Żammit, 2023;
that their peers would endorse its usage (Ibrahim et al., 2023). A Liu et al., 2024; Sallam and Mousa, 2024). Sex also appeared to play a
recently published study among marketing students in India indicated role in ChatGPT acceptance, with previous evidence suggesting that
that 309 out of 425 students were aware of ChatGPT (73%), with daily females may face more technical challenges and perceive greater risks

Frontiers in Education 09 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

TABLE 6 Regression analysis of the predictors influencing ChatGPT usage based on the TAME-ChatGPT constructs.

Model Coefficientsa
Unstandardized Standardized T statistic p value VIFc
Coefficients Coefficients
B SEb Beta
Constant 0.561 0.142 3.955 <0.001

Sex −0.003 0.023 −0.003 −0.109 0.913 1.068

Nationality −0.057 0.026 −0.055 −2.243 0.025 1.085

Cumulative GPA −0.017 0.012 −0.035 −1.434 0.152 1.070

Educational level −0.033 0.058 −0.014 −0.576 0.565 1.029

Perceived usefulness 0.398 0.027 0.460 14.643 <0.001 1.757

Behavior/cognitive factors 0.276 0.022 0.396 12.465 <0.001 1.798

Perceived risk of use 0.265 0.024 0.275 10.898 <0.001 1.139

Perceived ease of use 0.368 0.03 0.310 12.181 <0.001 1.158

a
Dependent variable: Overall TAME-ChatGPT usage score.
b
SE: Standard error.
c
VIF: Variance inflation factor.
Statistically significant p values are highlighted in bold style. Adjusted R2 = 0.769, SE = 0.23. ANOVA F statistic = 172.48, p value < 0.001.

TABLE 7 Regression analysis of the predictors influencing the attitude towards ChatGPT based on the TAME-ChatGPT constructs.

Model Coefficientsa
Unstandardized Standardized T statistic p value VIFc
coefficients coefficients
B SEb Beta
Constant 0.285 0.073 3.907 <0.001 1.009

Age 0.031 0.02 0.029 1.547 0.123 1.016

College 0.029 0.022 0.024 1.287 0.199 1.678

Perceived risk 0.418 0.023 0.442 18.212 <0.001 1.670

Anxiety 0.479 0.022 0.537 22.173 <0.001 1.073

Attitude to technology/ 0.413 0.018 0.438 22.539 <0.001 1.009

social influence
a
Dependent variable: Overall TAME-ChatGPT usage score.
b
SE: Standard error.
c
VIF: Variance inflation factor.
Statistically significant p values are highlighted in bold style. Adjusted R2 = 0.808, SE = 0.23; ANOVA F statistic = 460.12, p value < 0.001.

when using technology compared to males (Goswami and Dutta, technologies, likely influenced by a combination of cultural/societal
2016; Cai et al., 2017). The more positive attitude observed among norms and prior educational experiences as well as the attitudes
students from non-health disciplines can be attributed to their specific towards the perceived risk.
academic fields. University students in technology-related disciplines The propensity of postgraduate students to exhibit these
might be more inclined towards embracing new technologies such as characteristics could be related to their extended experience with a
ChatGPT which is influenced by both the curricular content and variety of technologies. Consequently, this extended exposure could
personal interests. This notion is supported by a study by Margaryan result in a more readiness to accept and engage with innovative
et al. (2011), which found that engineering students were more likely technological tools such as ChatGPT. Likewise, students in non-health
to use technology tools for various purposes compared to social disciplines especially in technology-related colleges might be more
work students. regularly exposed to emerging technologies. Consequently, this
The univariate analysis of the usage constructs of TAME-ChatGPT exposure would lead to increased familiarity with novel technologies
in this study revealed the following findings. The lower perceived risk and a lower level of perceived risks. Additionally, the previous
of ChatGPT use and higher agreement with behavior/cognitive academic experience with novel technologies among these students
factors, indicative of an instinctive impulse to utilize this novel could result in an innate readiness to engage with technological
technology, were associated with being male, Arab, enrolled in advancements such as ChatGPT.
non-health colleges, and having a lower GPA. These associations may The correlation of lower perceived risk from ChatGPT and more
point to a greater tendency among these groups to embrace new agreement with cognitive/behavioral factors with lower GPA

Frontiers in Education 10 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

categories was an interesting finding in this study. One possible of ChatGPT in learning has been demonstrated in a study involving
explanation is that university students with lower academic nursing students (Savellon et al., 2024).
performance are more inclined to experiment with novel technologies The significance of perceived risks in ChatGPT use, which includes
such as ChatGPT as a compensatory mechanism to improve their concerns about cybersecurity (Mijwil et al., 2023a), bias (Ray, 2023),
academic achievements. Another explanation could be related to the and inaccuracies (Borji, 2023; Sallam, 2023), was a critical determinant
higher propensity to explore innovative tools including ChatGPT as a of both attitude to ChatGPT and its usage in this study. This finding
result of the lower level of perceived constraints of academic rigor. In was consistent with a recent qualitative study which utilized the
all cases, this particular observation warrants further investigation to UTAUT model and highlighted the role of privacy concerns,
understand the underlying motivations and implications of ChatGPT performance expectancy, effort expectancy, social influence, and
adoption among university students with varying academic facilitating conditions in driving engagement with ChatGPT (Menon
performances which may require tailoring the adoption of ChatGPT and Shilpa, 2023). This suggests that university students’ collective
to the individual student needs. perceptions of both benefits and risks posed by technologies such as
In this study, the perceived usefulness of ChatGPT, behavioral/ ChatGPT play a key role to shape their engagement and adoption of
cognitive factors, perceived risk associated with ChatGPT usage, and this novel technology (Chan and Hu, 2023; Abdaljaleel et al., 2024).
the perceived ease of using ChatGPT were all significantly correlated In this study, multivariate regression analyses provided
with the overall ChatGPT usage score. Concerning the overall TAME- comprehensive insights into the predictors of ChatGPT usage, and
ChatGPT attitude score, lower perceived risk and anxiety were attitudes as modeled by the TAME-ChatGPT constructs. The complex
associated with more favorable attitudes towards ChatGPT, alongside nature of ChatGPT perceptions and adoption were determined by
a positive attitude towards technology and social influence. These several factors including the psycho-social determinants manifested in
findings contribute an additional evidence to the growing literature anxiety, behavior/cognitive and social influence constructs besides the
emphasizing the significance of various constructs in technology individual perception of usefulness, usability, and perception of risks.
acceptance assessment tools, such as the TAM and the Unified Theory Notably, the demographic and academic variables (e.g., age, sex, college,
of Acceptance and Use of Technology (UTAUT2) in the adoption of GPA) were not significant predictors of ChatGPT attitude or usage. This
ChatGPT in various settings (Foroughi et al., 2023; Habibi et al., 2023; result suggests that practical aspects of this novel technology such as the
Jo and Bang, 2023; Yilmaz et al., 2023). In turn, this can help to guide usefulness, user-friendly nature, and potential risks were more impactful
evidence-based strategies to govern ChatGPT use among other genAI in determining usage than demographic attributes. Thus, the primary
models in higher education (Veras et al., 2023; Grájeda et al., 2024). drivers of ChatGPT usage and attitudes are mainly rooted in students’
For example, a study among Polish university students using psycho-social predispositions. These findings suggest that if ChatGPT
UTAUT2 found that habit was the most influential factor on among other genAI models are to be integrated in higher education,
behavioral intention in the context of ChatGPT use, followed by there is a necessity for strategies to enhance the perceived usefulness and
performance expectancy and hedonic motivation (Strzelecki, 2023). ease of use for these models. Additionally, addressing the students’
Moreover, behavioral intention had the most substantial effect on anxiety and perceived risks is required for positive engagement with
ChatGPT usage behavior, followed by habit and facilitating conditions genAI technology in education. Nevertheless, the demographic and
(Strzelecki, 2023). Our results align with these findings, particularly academic variables can also be considered to achieve an intricate
regarding behavioral/cognitive factors, which are reflected in items understanding of attitude and use of ChatGPT among university students.
indicating previous use of similar tools and frequent utilization of Based on the results of this study, it is crucial for higher education
ChatGPT in university assignments. Performance expectations in our institutions, policymakers, and educators to formulate new
study were analogous to the perceived usefulness construct, while educational strategies to accommodate the transformative AI changes.
facilitating conditions are comparable to our perceived ease of use These strategies should highlight the utility and accessibility of genAI
construct highlighting recurrence of similar themes for the factors tools and proactively address the potential apprehensions that students
driving the acceptance of ChatGPT as an innovative tool in higher might encounter as shown recently by Oluwadiya et al. (2023). To
education (Gupta and Yang, 2024). The perceived ease of use is enhance the perception of ease and usefulness of genAI models,
particularly an important driver for the wide popularity of ChatGPT higher education institutions are advised to launch AI integration
with a user-friendly interface and little technical requirements (Shaikh initiatives that include raising awareness regarding the benefits of
et al., 2023; Albayati, 2024). In turn, this perceived ease of use would these tools in education (Ivanov et al., 2024). These initiatives should
render ChatGPT more appealing for students to try and also focus on providing training sessions, tutorials, and practical,
continue using. hands-on experiences for students and faculty alike to explore the full
In line with our findings, the importance of usefulness and ease of potential of genAI in education (Chiu, 2024). Furthermore, integrating
use has been shown in a recent study by Almogren et al. (2024) among these genAI tools into the curriculum can directly benefit learning and
a group of undergraduate and postgraduate university students. research by demonstrating their real-world applications (Sheikh Faisal
Additionally, the effectiveness of AI tools was an important predictor et al., 2024). These applications include generating content for study
of its acceptance and use among university students in a recent study materials, or simulating complex concepts, thereby enriching the
conducted in Malaysia and Pakistan (Dahri et al., 2024). Moreover, educational experience (Yu, 2024). The better usability of ChatGPT
effectiveness has been shown to positively influence ChatGPT usage compared to web-based tools in health education has been suggested
frequency as shown in a recent study in a different context by de in a recent research protocol by Veras et al. (2023). In another aspect,
Winter et al. (2024). The importance of usefulness has also been addressing psychological barriers such as anxiety and perceived risks
shown through its influence on user satisfaction in a recent study associated with genAI technology is essential and can be effectively
addressing AI chatbots user experience (Xing and Jiang, 2024). managed by shifting the cultural dynamics within educational
Furthermore, the central role of perceived usefulness and ease of use institutions (Yusuf et al., 2024). Programs designed to familiarize

Frontiers in Education 11 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

students and faculties with genAI tools can facilitate the integration of significant factors driving the attitude towards ChatGPT and its usage.
this novel technology into higher education and alleviate fears, similar These insights can help higher education institutions, policymakers,
to initiatives seen with the introduction of digital learning tools in the and educators to formulate clear initiatives and guidelines that would
past decade as reviewed recently by Fernández et al. (2023). help students in circumventing the ethical and practical aspects of
Although demographic and academic variables may not emerge genAI tool adoption in higher education.
as primary drivers in genAI adoption, their influence on technology
interaction should not be overlooked. Tailored educational practices
that consider these factors can significantly enhance genAI adoption Data availability statement
rates. By acknowledging these subtle differences, educational policies
can better accommodate a diverse students’ strata, ensuring that the The original contributions presented in the study are included in
benefits of genAI are accessible to all students, regardless of their the article/supplementary material, further inquiries can be directed
background (Sheikh Faisal et al., 2024). These strategic approaches can to the corresponding authors.
facilitate the adoption of genAI technologies and maximize their
potential to enrich learning experiences and outcomes. Subsequently,
this can help to prepare students to operate effectively in an Ethics statement
increasingly digital AI-driven world (George, 2023).
The results of this study highlighted several areas for future research This study was approved by the Institutional Review Board (IRB)
to enhance the collective understanding of genAI effects in different at Gulf Medical University (Reference number: IRB-COD-FAC-49-
educational contexts. For example, longitudinal studies are essential to APRIL-2023). Obtaining the informed consent to participation was
assess the long-term impact of genAI on learning outcomes and ensured by the inclusion of a mandatory item at the beginning of the
experiences, tracking changes in students’ perceptions and academic electronic survey to explicitly indicate consent for participation. The
performance over extended periods. Additionally, experimental designs studies were conducted in accordance with the local legislation and
like randomized controlled trials (RCTs) are recommended to establish institutional requirements. The participants provided their written
causal links between genAI usage and educational outcomes, building informed consent to participate in this study.
on recent protocols such as the one conceived by Veras et al. (2023).
Investigating genAI influence across different academic disciplines
would also be beneficial, which would help to develop tailored Author contributions
integration strategies that address the specific needs and challenges for
various academic disciplines. Moreover, considering the significant role MS: Conceptualization, Data curation, Formal analysis,
of cultural context in technology adoption, cross-cultural studies could Investigation, Methodology, Resources, Software, Supervision,
examine how different settings influence genAI acceptance and Validation, Writing – original draft, Writing – review & editing. WE:
effectiveness. Lastly, with ongoing concerns about the ethical Data curation, Investigation, Methodology, Writing – review &
implications and risks of genAI, further research should focus on these editing. MA-S: Data curation, Investigation, Methodology, Writing –
areas, particularly privacy, data security, and bias, to ensure responsible review & editing. MB: Data curation, Investigation, Methodology,
and ethical use of genAI for students and educational institutions. Writing – review & editing. SE: Data curation, Investigation,
Finally, it is important to consider the findings of this study in Methodology, Writing – review & editing. WG: Data curation,
light of several limitations as follows. The convenience sampling Investigation, Methodology, Writing – review & editing. NA: Data
approach utilized with the inherent selection bias could limit the curation, Investigation, Methodology, Writing – review & editing. SH:
representativeness of the sample and generalizability of the findings. Data curation, Investigation, Methodology, Writing – review &
The selection bias is also expected considering the electronic editing. DM: Data curation, Investigation, Methodology,
distribution of the survey among students. An element of bias should Project administration, Supervision, Validation, Writing – review &
be considered as well in light of more inclination of the students who editing.
previously engaged with ChatGPT or heard of it to participate in the
study and express their opinions. Finally, the reliance on self-reported
data could result in self-reporting bias. Funding
The author(s) declare that no financial support was received for
5 Conclusion the research, authorship, and/or publication of this article.

This study elucidated the determinants of ChatGPT adoption

among university students in the UAE. Addressing these factors could Conflict of interest
help to exploit ChatGPT potential for better learning experience and
to help equip university students to responsibly use the current and The authors declare that the research was conducted in the
future technological innovations. Students’ familiarity with ChatGPT absence of any commercial or financial relationships that could
can provide an opportunity for genAI integration in higher education be construed as a potential conflict of interest.
curricula and teaching methods. The author(s) declared that they were an editorial board member
The study highlighted the central role of individual and psycho- of Frontiers, at the time of submission. This had no impact on the peer
social factors as modeled in the TAME-ChatGPT constructs as review process and the final decision.

Frontiers in Education 12 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

Publisher’s note or those of the publisher, the editors and the reviewers. Any
product that may be evaluated in this article, or claim that may
All claims expressed in this article are solely those of the authors be made by its manufacturer, is not guaranteed or endorsed by the
and do not necessarily represent those of their affiliated organizations, publisher.

References
Abdaljaleel, M., Barakat, M., Alsanafi, M., Salim, N. A., Abazid, H., Malaeb, D., et al. de Winter, J., Dodou, D., and Eisma, Y. B. (2024). Personality and acceptance as
(2024). A multinational study on the factors influencing university students’ attitudes predictors of ChatGPT use. Discov. Psychol. 4:57. doi: 10.1007/s44202-024-00161-2
and usage of ChatGPT. Sci. Rep. 14:1983. doi: 10.1038/s41598-024-52549-8
de Winter, J. C. F., Dodou, D., and Stienen, A. H. A. (2023). ChatGPT in education:
Albayati, H. (2024). Investigating undergraduate students' perceptions and awareness empowering educators through methods for recognition and assessment. Informatics
of using ChatGPT as a regular assistance tool: a user acceptance perspective study. 10:87. doi: 10.3390/informatics10040087
Comput. Educ.: Artif. Intell. 6:100203. doi: 10.1016/j.caeai.2024.100203
Dempere, J., Modugu, K., Hesham, A., and Ramasamy, L. K. (2023). The impact of
Alkhaldi, F. K., and Altaei, S. (2021). “Emirates leading experience in employing ChatGPT on higher education. Front. Educ. 8:1206936. doi: 10.3389/feduc.2023.1206936
artificial intelligence” in The fourth industrial revolution: Implementation of artificial
Epitools – Epidemiological Calculators (2024) Sample size to estimate a proportion
intelligence for growing business success. eds. A. Hamdan, A. E. Hassanien, A. Razzaque
or apparent prevalence with specified precision. Available at: https://epitools.ausvet.com.
and B. Alareeni (Cham: Springer International Publishing), 241–251.
au/oneproportion (Accessed January 26, 2024)
Almogren, A. S., Al-Rahmi, W. M., and Dahri, N. A. (2024). Exploring factors
Farina, M., and Lavazza, A. (2023). ChatGPT in society: emerging issues. Front. Artif.
influencing the acceptance of ChatGPT in higher education: a smart education
Intell. 6:1130913. doi: 10.3389/frai.2023.1130913
perspective. Heliyon 10:e31887. doi: 10.1016/j.heliyon.2024.e31887
Fernández, A., Gómez, B., Binjaku, K., and Meçe, E. K. (2023). Digital transformation
Ansari, A. N., Ahmad, S., and Bhutta, S. M. (2023). Mapping the global evidence
initiatives in higher education institutions: a multivocal literature review. Educ. Inf.
around the use of ChatGPT in higher education: a systematic scoping review. Educ. Inf.
Technol. 28, 12351–12382. doi: 10.1007/s10639-022-11544-0
Technol. 29, 11281–11321. doi: 10.1007/s10639-023-12223-4
Foroughi, B., Senali, M. G., Iranmanesh, M., Khanfar, A., Ghobakhloo, M.,
Badry, F. (2019). “Expanding the UAE’s higher education horizon: path toward a
Annamalai, N., et al. (2023). Determinants of intention to use ChatGPT for educational
sustainable future” in Education in the United Arab Emirates: Innovation and
purposes: findings from PLS-SEM and fsQCA. Int. J. Hum.-Comput. Interact., 1–20. doi:
transformation. ed. K. Gallagher (Singapore: Springer Singapore), 57–73.
10.1080/10447318.2023.2226495
Bagozzi, R. P., Davis, F. D., and Warshaw, P. R. (1992). Development and test of a
Fütterer, T., Fischer, C., Alekseeva, A., Chen, X., Tate, T., Warschauer, M., et al. (2023).
theory of technological learning and usage. Hum. Relat. 45, 659–686. doi:
ChatGPT in education: global reactions to AI innovations. Sci. Rep. 13:15310. doi:
10.1177/001872679204500702
10.1038/s41598-023-42227-6
Bandi, A., Adapa, P. V., and Kuchi, Y. E. (2023). The power of generative AI: a review
George, A. S. (2023). Preparing students for an AI-driven world: rethinking
of requirements, models, input–output formats, evaluation metrics, and challenges.
curriculum and pedagogy in the age of artificial intelligence. Partners Univers. Innovative
Future Internet 15:260. doi: 10.3390/fi15080260
Res. Publ. 1, 112–136. doi: 10.5281/zenodo.10245675
Barakat, M., Salim, N. A., and Sallam, M. (2024). Perspectives of university educators
Goswami, A., and Dutta, S. (2016). Gender differences in technology
regarding ChatGPT: a validation study based on the technology acceptance model. Res.
usage—a literature review. Open J. Bus. Manage. 4, 51–59. doi: 10.4236/ojbm.
Sq. doi: 10.21203/rs.3.rs-3919524/v1
2016.41006
Barwise, A. K., Curtis, S., Diedrich, D. A., and Pickering, B. W. (2024). Using artificial
Grájeda, A., Burgos, J., Córdova, P., and Sanjinés, A. (2024). Assessing student-
intelligence to promote equitable care for inpatients with language barriers and complex
perceived impact of using artificial intelligence tools: construction of a synthetic index
medical needs: clinical stakeholder perspectives. J. Am. Med. Inf. Assoc. 31, 611–621. doi:
of application in higher education. Cogent Educ. 11:2287917. doi: 10.1080/
10.1093/jamia/ocad224
2331186X.2023.2287917
Borji, A. (2023). A categorical archive of ChatGPT failures. Res. Sq. doi: 10.21203/
Gulati, A., Saini, H., Singh, S., and Kumar, V. (2024). Enhancing learning potential:
rs.3.rs-2895792/v1
investigating marketing students’ behavioral intentions to adopt ChatGPT. Mark. Educ.
Bukartaite, R., and Hooper, D. (2023). Automation, artificial intelligence and future skills Rev. 1-34, 1–34. doi: 10.1080/10528008.2023.2300139
needs: an Irish perspective. Eur. J. Train. Dev. 47, 163–185. doi: 10.1108/EJTD-03-2023-0045
Gupta, V. P. (2024). “AI-driven skill development: bridging students with industry 5.0”
Cai, Z., Fan, X., and Du, J. (2017). Gender and attitudes toward technology use: a in Infrastructure possibilities and human-centered approaches with industry 5.0. eds.
meta-analysis. Comput. Educ. 105, 1–13. doi: 10.1016/j.compedu.2016.11.003 M. A. Khan, R. Khan, P. Praveen, A. R. Verma and M. K. Panda (Hershey, PA, USA: IGI
Global), 56–68.
Caleb, K., Lenny, W., Dazzy, I., and Azra, E. (2023). The impact of AI on teaching and
learning. London J. Soc. Sci. 6, 124–129. doi: 10.31039/ljss.2023.6.111 Gupta, V., and Yang, H. (2024). Generative artificial intelligence (AI) technology
adoption model for entrepreneurs: case of ChatGPT. Internet Ref. Serv. Q. 28, 223–242.
Chan, C. K. Y. (2023). A comprehensive AI policy education framework for university
doi: 10.1080/10875301.2023.2300114
teaching and learning. Int. J. Educ. Technol. High. Educ. 20:38. doi: 10.1186/
s41239-023-00408-3 Habibi, A., Muhaimin, M., Danibao, B. K., Wibowo, Y. G., Wahyuni, S., and Octavia, A.
(2023). ChatGPT in higher education learning: acceptance and use. Comput. Educ.: Artif.
Chan, C. K. Y., and Hu, W. (2023). Students’ voices on generative AI: perceptions,
Intell. 5:100190. doi: 10.1016/j.caeai.2023.100190
benefits, and challenges in higher education. Int. J. Educ. Technol. High. Educ. 20:43. doi:
10.1186/s41239-023-00411-8 Haleem, A., Javaid, M., Qadri, M. A., and Suman, R. (2022). Understanding the role
of digital technologies in education: a review. Sustainable Oper. Comput. 3, 275–285. doi:
Chen, T. J. (2023). ChatGPT and other artificial intelligence applications speed up
10.1016/j.susoc.2022.05.004
scientific writing. J. Chin. Med. Assoc. 86, 351–353. doi: 10.1097/jcma.0000000000000900
Ibrahim, H., Liu, F., Asim, R., Battu, B., Benabderrahmane, S., Alhafni, B., et al. (2023).
Chiu, T. K. F. (2024). Future research recommendations for transforming higher
Perception, performance, and detectability of conversational artificial intelligence across
education with generative AI. Comput. Educ.: Artif. Intell. 6:100197. doi: 10.1016/j.
32 university courses. Sci. Rep. 13:12187. doi: 10.1038/s41598-023-38964-3
caeai.2023.100197
Ivanov, S., Soliman, M., Tuomi, A., Alkathiri, N. A., and Al-Alawi, A. N. (2024).
Criollo-C, S., Govea, J., Játiva, W., Pierrottet, J., Guerrero-Arias, A.,
Drivers of generative AI adoption in higher education through the lens of the theory of
Jaramillo-Alcázar, Á., et al. (2023). Towards the integration of emerging technologies as
planned behaviour. Technol. Soc. 77:102521. doi: 10.1016/j.techsoc.2024.102521
support for the teaching and learning model in higher education. Sustain. For. 15:6055.
doi: 10.3390/su15076055 Jo, H., and Bang, Y. (2023). Analyzing ChatGPT adoption drivers with the TOEK
framework. Sci. Rep. 13:22606. doi: 10.1038/s41598-023-49710-0
Dahabreh, F. (2023) The continued usage of artificial intelligence in the
United Arab Emirates public sector organisations: An extended information system Johnson, W. L. (2023). How to harness generative AI to accelerate human learning.
success model. Northumbria University. Int. J. Artif. Intell. Educ. doi: 10.1007/s40593-023-00367-w
Dahri, N. A., Yahaya, N., Al-Rahmi, W. M., Vighio, M. S., Alblehai, F., Soomro, R. B., Kamalov, F., Santandreu Calonge, D., and Gurrib, I. (2023). New era of artificial
et al. (2024). Investigating AI-based academic support acceptance and its impact on intelligence in education: towards a sustainable multifaceted revolution. Sustain. For.
students’ performance in Malaysian and Pakistani higher education institutions. Educ. 15:12451. doi: 10.3390/su151612451
Inf. Technol. doi: 10.1007/s10639-024-12599-x
Karen, W., James, H., and Kathryn, A. (2023). “Perceptions and barriers to adopting
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance artificial intelligence in K-12 education: a survey of educators in fifty states” in
of information technology. MIS Q. 13, 319–340. doi: 10.2307/249008 Reimagining education. ed. M. Sharon (Rijeka: IntechOpen).

Frontiers in Education 13 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

King, S., and Prasetyo, J. (2023). Assessing generative A.I. Through the lens of the 2023 Sallam, M., Al-Farajat, A., and Egger, J. (2024). Envisioning the future of ChatGPT in
Gartner hype cycle for emerging technologies: a collaborative autoethnography. Front. healthcare: insights and recommendations from a systematic identification of influential
Educ. 8:1300391. doi: 10.3389/feduc.2023.1300391 research and a call for papers. Jordan Med. J. 58, 236–249. doi: 10.35516/jmj.v58i1.2285
Kitsara, I. (2022). “Artificial intelligence and the digital divide: from an innovation Sallam, M., and Mousa, D. (2024). Evaluating ChatGPT performance in Arabic dialects: a
perspective” in Platforms and artificial intelligence: The next generation of competences. comparative study showing defects in responding to Jordanian and Tunisian general health
ed. A. Bounfour (Cham: Springer International Publishing), 245–265. prompts. Mesopotamian J. Artif. Intell. Healthcare 2024, 1–7. doi: 10.58496/MJAIH/2024/001
Kıyak, Y. S., and Emekli, E. (2024). ChatGPT prompts for generating multiple-choice Sallam, M., Salim, N. A., Barakat, M., Al-Mahzoum, K., Al-Tammemi, A. B.,
questions in medical education and evidence on their validity: a literature review. Malaeb, D., et al. (2023a). Assessing health Students' attitudes and usage of ChatGPT in
Postgrad. Med. J.:qgae065. doi: 10.1093/postmj/qgae065 Jordan: validation study. JMIR Med. Educ. 9:e48254. doi: 10.2196/48254
Kurtz, G., Amzalag, M., Shaked, N., Zaguri, Y., Kohen-Vacs, D., Gal, E., et al. (2024). Sallam, M., Salim, N. A., Barakat, M., and Al-Tammemi, A. B. (2023b). ChatGPT
Strategies for integrating generative AI into higher education: navigating challenges and applications in medical, dental, pharmacy, and public health education: a descriptive study
leveraging opportunities. Educ. Sci. 14:503. doi: 10.3390/educsci14050503 highlighting the advantages and limitations. Narra J. 3:e103. doi: 10.52225/narra.v3i1.103
Lim, W. M., Gunasekara, A., Pallant, J. L., Pallant, J. I., and Pechenkina, E. (2023). Savellon, M. U., Baybayan, S. B., and Asiri, M. S. (2024). Learning satisfaction on the
Generative AI and the future of education: Ragnarök or reformation? A paradoxical use of chatgpt among nursing students in selected higher education institutions in Sulu.
perspective from management educators. Int. J. Manage. Educ. 21:100790. doi: 10.1016/j. J. Educ. Acad. Settings 1, 1–16. doi: 10.62596/t3wgsm55
ijme.2023.100790
Science/AAAS Custom Publishing Office (2023) The world’s first AI university pairs
Liu, X., Wu, J., Shao, A., Shen, W., Ye, P., Wang, Y., et al. (2024). Uncovering language machine learning’s great power with great responsibility. Available at: https://www.
disparity of ChatGPT on retinal vascular disease classification: cross-sectional study. J. science.org/content/resource/worlds-first-ai-university-pairs-machine-learnings-great-
Med. Internet Res. 26:e51926. doi: 10.2196/51926 power-great-responsibility (Accessed June 23, 2024)
Marangunić, N., and Granić, A. (2015). Technology acceptance model: a literature review Shaikh, S., Yayilgan, S. Y., Klimova, B., and Pikhart, M. (2023). Assessing the usability
from 1986 to 2013. Univ. Access Inf. Soc. 14, 81–95. doi: 10.1007/s10209-014-0348-1 of ChatGPT for formal English language learning. Eur. J. Investig. Health Psychol. 13,
Margaryan, A., Littlejohn, A., and Vojt, G. (2011). Are digital natives a myth or reality? 1937–1960. doi: 10.3390/ejihpe13090140
University students’ use of digital technologies. Comput. Educ. 56, 429–440. doi: Sheikh Faisal, R., Nghia, D.-T., and Niels, P. (2024). “Generative AI in education:
10.1016/j.compedu.2010.09.004 technical foundations, applications, and challenges” in Artificial intelligence for quality
Menon, D., and Shilpa, K. (2023). “Chatting with ChatGPT”: analyzing the factors education. ed. K. Seifedine (Rijeka: IntechOpen).
influencing users' intention to use the open AI's ChatGPT using the UTAUT model. Shwedeh, F., Salloum, S. A., Aburayya, A., Fatin, B., Elbadawi, M. A., Al Ghurabli, Z.,
Heliyon 9:e20962. doi: 10.1016/j.heliyon.2023.e20962 et al. (2024). “AI adoption and educational sustainability in higher education in the
Michel-Villarreal, R., Vilalta-Perdomo, E., Salinas-Navarro, D. E., Thierry-Aguilera, R., UAE” in Artificial intelligence in education: The power and dangers of ChatGPT in the
and Gerardou, F. S. (2023). Challenges and opportunities of generative AI for higher classroom. eds. A. Al-Marzouqi, S. A. Salloum, M. Al-Saidat, A. Aburayya and B. Gupta
education as explained by ChatGPT. Educ. Sci. 13:856. doi: 10.3390/educsci13090856 (Switzerland, Cham: Springer Nature), 201–229.

Mijwil, M. M., Abdulrhman, S. H., Abttan, R. A., Faieq, A. K., and Alkhazraji, A. Strzelecki, A. (2023). Students’ acceptance of ChatGPT in higher education: an
(2023b). Artificial intelligence applications in English language teaching: a short survey. extended unified theory of acceptance and use of technology. Innov. High. Educ. 49,
Asian J. Appl. Sci. 10. doi: 10.24203/ajas.v10i6.7111 223–245. doi: 10.1007/s10755-023-09686-1

Mijwil, M., Ali, G., Sadıkoğlu, E., Guma, A., and Emre, S. (2023). The evolving role of Tavakol, M., and Dennick, R. (2011). Making sense of Cronbach's alpha. Int. J. Med.
artificial intelligence in the future of distance learning: exploring the next frontier. Educ. 2, 53–55. doi: 10.5116/ijme.4dfb.8dfd
Mesopotamian J. Comput. Sci. 2023, 98–105. doi: 10.58496/MJCSC/2023/012 Tayan, O., Hassan, A., Khankan, K., and Askool, S. (2024). Considerations for
Mijwil, M. M., Aljanabi, M., and Ali, A. H. (2023a). ChatGPT: exploring the role of adapting higher education technology courses for AI large language models: a critical
cybersecurity in the protection of medical information. Mesopotamian J. CyberSecur. review of the impact of ChatGPT. Mach. Learn. Appl. 15:100513. doi: 10.1016/j.
2023, 18–21. doi: 10.58496/MJCS/2023/004 mlwa.2023.100513

Ng, D. T. K., Leung, J. K. L., Su, J., Ng, R. C. W., and Chu, S. K. W. (2023). Teachers’ AI The UAE Ministry of Education (2024) Open Data. Available at: https://www.moe.gov.
digital competencies and twenty-first century skills in the post-pandemic world. Educ. ae/En/OpenData/pages/home.aspx (Accessed January 26, 2024)
Technol. Res. Dev. 71, 137–161. doi: 10.1007/s11423-023-10203-6 Veras, M., Dyer, J. O., Rooney, M., Barros Silva, P. G., Rutherford, D., and Kairy, D.
Ogunleye, B., Zakariyyah, K. I., Ajao, O., Olayinka, O., and Sharma, H. (2024). A (2023). Usability and efficacy of artificial intelligence Chatbots (ChatGPT) for health
systematic review of generative AI for teaching and learning practice. Educ. Sci. 14:636. sciences students: protocol for a crossover randomized controlled trial. JMIR Res. Protoc.
doi: 10.3390/educsci14060636 12:e51873. doi: 10.2196/51873
Okoye, K., Hussein, H., Arrona-Palacios, A., Quintero, H. N., Ortega, L. O. P., von Garrel, J., and Mayer, J. (2023). Artificial intelligence in studies—use of ChatGPT
Sanchez, A. L., et al. (2023). Impact of digital technologies upon teaching and learning and AI-based tools among students in Germany. Humanit. Soc. Sci. Commun. 10:799.
in higher education in Latin America: an outlook on the reach, barriers, and bottlenecks. doi: 10.1057/s41599-023-02304-7
Educ. Inf. Technol. 28, 2291–2360. doi: 10.1007/s10639-022-11214-1 Weidener, L., and Fischer, M. (2024). Artificial intelligence in medicine: cross-
Oluwadiya, K. S., Adeoti, A. O., Agodirin, S. O., Nottidge, T. E., Usman, M. I., sectional study among medical students on application, education, and ethical aspects.
Gali, M. B., et al. (2023). Exploring artificial intelligence in the Nigerian medical JMIR Med. Educ. 10:e51247. doi: 10.2196/51247
educational space: an online cross-sectional study of perceptions, risks and benefits Williams, R. T. (2024). The ethical implications of using generative chatbots in higher
among students and lecturers from ten universities. Niger. Postgrad. Med. J. 30, 285–292. education. Front. Educ. 8:1331607. doi: 10.3389/feduc.2023.1331607
doi: 10.4103/npmj.npmj_186_23
Xing, J., and Jiang, Q. (2024). Factors influencing user experience in AI chat systems
Ouyang, F., and Jiao, P. (2021). Artificial intelligence in education: the three paradigms. – a satisfaction study based on factor analysis and linear regression. Kybernetes. doi:
Comput. Educ.: Artif. Intell. 2:100020. doi: 10.1016/j.caeai.2021.100020 10.1108/K-10-2023-2237
Ragnedda, M., and Muschert, G. (2013). The digital divide: the internet and social Yilmaz, F. G. K., Yilmaz, R., and Ceylan, M. (2023). Generative artificial intelligence
inequality in international perspective. 1st Edn. London: Routledge. acceptance scale: a validity and reliability study. Int. J. Hum.-Comput. Interact., 1–13.
Ray, P. P. (2023). ChatGPT: a comprehensive review on background, applications, key doi: 10.1080/10447318.2023.2288730
challenges, bias, ethics, limitations and future scope. Internet Things Cyber-Phys. Syst. 3, Yu, H. (2024). The application and challenges of ChatGPT in educational
121–154. doi: 10.1016/j.iotcps.2023.04.003 transformation: new demands for teachers' roles. Heliyon 10:e24289. doi: 10.1016/j.
Salazar, L. R., Peeples, S. F., and Brooks, M. E. (2024). “Generative AI ethical heliyon.2024.e24289
considerations and discriminatory biases on diverse students within the classroom” in Yu, H., and Guo, Y. (2023). Generative artificial intelligence empowers educational
The role of generative AI in the communication classroom. eds. S. Elmoudden and J. S. reform: current status, issues, and prospects. Front. Educ. 8:1183162. doi: 10.3389/
Wrench (Hershey, PA, USA: IGI Global), 191–213. feduc.2023.1183162
Salinas-Navarro, D. E., Vilalta-Perdomo, E., Michel-Villarreal, R., and Montesinos, L.
Yusuf, A., Pervin, N., and Román-González, M. (2024). Generative AI and the future
(2024a). Designing experiential learning activities with generative artificial intelligence tools
of higher education: a threat to academic integrity or reformation? Evidence from
for authentic assessment. Interact. Technol. Smart Educ. doi: 10.1108/ITSE-12-2023-0236
multicultural perspectives. Int. J. Educ. Technol. High. Educ. 21:21. doi: 10.1186/
Salinas-Navarro, D. E., Vilalta-Perdomo, E., Michel-Villarreal, R., and Montesinos, L. s41239-024-00453-6
(2024b). Using generative artificial intelligence tools to explain and enhance experiential
Żammit, J. (2023). “Harnessing the power of ChatGPT for mastering the Maltese
learning for authentic assessment. Educ. Sci. 14:83. doi: 10.3390/educsci14010083
language: a journey of breaking barriers and charting new paths” in Machine intelligence
Sallam, M. (2023). ChatGPT utility in healthcare education, research, and practice: for smart applications: Opportunities and risks. eds. A. Adadi and S. Motahhir
systematic review on the promising perspectives and valid concerns. Healthcare (Basel) (Switzerland, Cham: Springer Nature), 161–178.
11:887. doi: 10.3390/healthcare11060887
Zarifhonarvar, A. (2023). Economics of ChatGPT: a labor market view on the
Sallam, M. (2024). Bibliometric top ten healthcare-related ChatGPT publications in occupational impact of artificial intelligence. J. Electron. Bus. Digit. Econ. 3, 100–116.
the first ChatGPT anniversary. Narra J. 4:e917. doi: 10.52225/narra.v4i2.917 doi: 10.1108/JEBDE-10-2023-0021

Frontiers in Education 14 frontiersin.org

Sallam et al. 10.3389/feduc.2024.1414758

Appendix A1
Table A1.
TABLE A1 Full TAME-ChatGPT items.

Usage scale

A. Perceived usefulness

1. ChatGPT helps me to save time when searching for information

2. For me, ChatGPT is a reliable source of accurate information

3. I recommend ChatGPT to my colleagues to facilitate their academic duties

4. ChatGPT is more useful than other sources of information that I have used previously

5. I appreciate the accuracy and reliability of the information provided by ChatGPT

6. I believe that using ChatGPT can save time and effort in my university assignments and duties

B. Behavior/cognitive factors

7. I have used tools or techniques similar to ChatGPT in the past

8. I spontaneously find myself using ChatGPT when I need information for my university assignments and duties

9. I often use ChatGPT as a source of information in my university assignments and duties

C. Perceived risk of use REVERSED SCORE

10. I am concerned that using ChatGPT would get me accused of plagiarism

11. I am concerned about the potential security risks of using ChatGPT

12. I think that relying on technology like ChatGPT can disrupt my critical thinking skills

D. Perceived ease of use

13. It does not take a long time to learn how to use ChatGPT

14. ChatGPT does not require extensive technical knowledge

Attitude scale

A. Perceived risk REVERSED SCORE

1. I am concerned about the reliability of the information provided by ChatGPT

2. I am concerned that using ChatGPT would get me accused of plagiarism

3. I am concerned about the potential security risks of using ChatGPT

4. I am afraid that the use of the ChatGPT would be a violation of academic and university policies

5. I am concerned about the potential privacy risks that might be associated with using ChatGPT

B. Anxiety REVERSED SCORE

6. I am afraid of relying too much on ChatGPT and not developing my critical thinking skills

7. I am afraid of becoming too dependent on technology like ChatGPT

8. I am afraid that using ChatGPT would result in a lack of originality in my university assignments and duties

C. Technology/social influence

9. I am enthusiastic about using technology such as ChatGPT for learning and research

10. I believe technology such as ChatGPT is an important tool for academic success

11. I think that technology like ChatGPT is attractive and fun to use

12. I am always keen to learn about new technologies like ChatGPT

13. I trust the opinions of my friends or colleagues about using ChatGPT

Frontiers in Education 15 frontiersin.org

View publication stats