Academic Integrity and Students’ Ethical Use of ChatGPT in Higher Education
Open accessThe DOI matches its Crossref record; 151 of the 152 references it cites were found in a registry or an index; 19 of the 51 sources its claims cite were checked for support.
Abstract
Aim/Purpose: To examine how ethical awareness, cognitive appraisal (trust and perceived usefulness), digital competence, academic performance, and gender influence university students’ ethical use of ChatGPT and academic integrity. Background: This study explores how university students’ use of ChatGPT influences academic integrity in higher education. It responds to emerging integrity challenges posed by generative AI by empirically testing a model that links transparency, plagiarism avoidance, bias awareness, and responsible use to academic integrity outcomes across universities in the Gulf region. Methodology: PLS-SEM analysis of survey data from 318 students across five Gulf-region universities; tests direct effects of four ethical variables, mediation by trust in AI and perceived usefulness, and moderation by digital literacy and CGPA. Contribution: This study provides empirical evidence that core ethical-use dimensions significantly enhance academic integrity; clarifies the mediating roles of trust/usefulness and the moderating roles of digital literacy/CGPA; and documents gender and discipline differences in usage. Findings: All four ethical variables positively and significantly predict academic integrity: Trust in AI and perceived usefulness act as partial mediators. Digital literacy and CGPA significantly moderate several relationships. High-performing and senior students report more frequent and effective use of ChatGPT; gender and discipline differences are evident. Recommendations for Practitioners: Embed digital literacy and ethical-AI training (transparency, anti-plagiarism, bias awareness, responsible use) into curricula; implement inclusive AI policies; and guide trust calibration and verification workflows to support responsible use. Recommendation for Researchers: Extend the model across countries and disciplines; examine longitudinal effects; test additional mediators (e.g., AI anxiety, institutional policy clarity) and moderators (e.g., year of study, assessment type); compare alternative SEM and causal designs. Impact on Society: Promotes responsible AI adoption that safeguards academic ethics, supports equitable student outcomes, and informs policy for trustworthy AI use in higher education ecosystems. Future Research: Conduct multi-institutional replications, experimental interventions on ethics/digital literacy training, and studies of assessment design that balance AI use with integrity (e.g., oral/ authentic assessments).
Overview
CoChat Summary
TL;DR
This study tests a PLS-SEM model of how ethical-use dimensions—transparency, plagiarism avoidance, bias awareness, and responsible use—relate to academic integrity among university students using ChatGPT. Across 318 students in five Gulf-region universities, all four ethical dimensions positively predicted academic integrity, with trust in AI and perceived usefulness as partial mediators and digital literacy/CGPA as significant moderators.
Key findings
- All four direct ethical predictors significantly and positively predicted academic integrity: transparency (β = 0.28, t = 4.21, p < 0.001), responsible use (β = 0.26, t = 4.00, p < 0.001), plagiarism avoidance (β = 0.22, t = 3.68, p < 0.001), and bias awareness (β = 0.19, t = 3.01, p < 0.001); R² for academic integrity was 0.48.
- Trust in AI and perceived usefulness partially mediated these relationships: indirect effects ranged from β = 0.099 to 0.13, all significant, and R² for academic integrity increased to 0.55 with mediators included.
- Digital literacy and CGPA significantly moderated several relationships: transparency × digital literacy (β = 0.14, p = 0.004), bias awareness × digital literacy (β = 0.11, p = 0.035), plagiarism avoidance × CGPA (β = 0.13, p = 0.008), and responsible use × CGPA (β = 0.16, p = 0.001); R² with moderators was 0.61.
- High-performing and senior students reported more frequent and effective ChatGPT use, while gender differences were evident: male students reported higher usage across all CGPA ranges, especially programming and problem-solving, whereas female students showed lower engagement.
Methods
- Quantitative cross-sectional survey design; final valid sample of 318 undergraduate university students in the United Arab Emirates, recruited via non-probability convenience sampling from five public and private universities in Dubai, Sharjah, and Ajman.
- Instrument development began with 36 items, refined to 20 after expert review; a pilot with 35 students was excluded from final data collection; all construct items used a 5-point Likert scale.
- Data were analyzed with Partial Least Squares Structural Equation Modeling (PLS-SEM), testing direct effects, mediation by trust in AI and perceived usefulness, and moderation by digital literacy and CGPA.
Limitations
- The paper states that the study was conducted in five universities in the Gulf region, limiting generalizability to other cultural and institutional contexts; self-reported survey data may be affected by social desirability and recall bias and do not capture actual AI usage patterns.
- The paper also states sample imbalances in gender and academic disciplines, with more male students and business-related majors, and that the study focused exclusively on ChatGPT, excluding other generative AI tools.
- Evident from the text: the cross-sectional design and single data source restrict causal inference and methodological triangulation, and convenience sampling may overrepresent digitally engaged or motivated students.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures






Verification
Claims the paper makes
12 Valid · 0 Not valid · 17 Not checked · 140 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Although existing studies examine AI adoption and traditional academic misconduct, they offer limited empirical insight into how students conceptualize ethical responsibility when using generative AI tools.”
§ ACADEMIC INTEGRITY AND STUDENTS' ETHICAL USE OF CHATGPT IN HIGHER EDUCATION > RESEARCH GAP
What the sources say
- Could not verifycites Guerrero-Dib et al., 2020 sourceExperimental — accuracy not yet measured
The Methodology and Discussion examine academic misconduct, students’ perceptions of dishonesty, and ethical behaviour. Neither those passages nor searches for generative AI and AI adoption address how students conceptualize ethical responsibility when using generative AI tools, or establish a gap in empirical research on that question. (LLM, conf=0.76)
- Could not verifycites Ramdani, 2018 sourceExperimental — accuracy not yet measured
The Introduction (p. 89) identifies a gap in how academic integrity had been measured, and the Introduction discusses traditional plagiarism. Neither passage addresses AI adoption or whether studies offer limited empirical insight into students’ ethical responsibility when using generative AI tools. Searches for those concepts found no relevant passage in our copy. (LLM, conf=0.76)
- Partly supportedcites Zawacki-Richter et al., 2019 sourceExperimental — accuracy not yet measured
The review reports a dramatic lack of critical reflection on the ethical implications of AI in higher education (discussion section; page unavailable), supporting the broad claim that ethical questions receive limited attention. It does not establish a lack of empirical insight into how students conceptualize ethical responsibility when using generative AI tools; its reported gap concerns AI applications more broadly. (LLM, conf=0.85)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Could not verify
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Furthermore, theoretical discussions often lack practical, evidence-based strategies to guide the ethical use of AI in educational contexts (Guerrero-Dib et al., 2020;Ramdani, 2018;Zawacki-Richter et al., 2019).”
§ ACADEMIC INTEGRITY AND STUDENTS' ETHICAL USE OF CHATGPT IN HIGHER EDUCATION > RESEARCH GAP
What the sources say
- Partly supportedcites Guerrero-Dib et al., 2020 sourceExperimental — accuracy not yet measured
The Discussion says teaching academic integrity only in a theoretical or conceptual way is insufficient and emphasizes learning through real contexts and practices. It does not discuss ethical AI use or establish that theoretical discussions often lack practical, evidence-based strategies for it. (LLM, conf=0.82)
- Could not verifycites Ramdani, 2018 sourceExperimental — accuracy not yet measured
The Introduction (p. 89) discusses plagiarism and academic integrity, while the Design and Discussion sections describe developing an integrity scale from theory and field data. A search for AI, ethical AI use, and the claimed lack of practical evidence-based strategies found no passage addressing that point. (LLM, conf=0.76)
- Partly supportedcites Zawacki-Richter et al., 2019 sourceExperimental — accuracy not yet measured
The review’s discussion reports a dramatic lack of critical reflection on the ethical risks of AI in higher education and calls for more research on integrating AI into practice. It also notes a shortage of implementation and impact studies. However, it does not specifically find that theoretical discussions lack practical, evidence-based strategies for ethical AI use. (LLM, conf=0.82)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Language models such as ChatGPT have been shown to reproduce stereotypes and narrow demographic representations, reinforcing the need for inclusive training practices and governance mechanisms (Gozalo-Brizuela & Garrido-Merchan, 2023;Zhou et al., 2023).”
§ ACADEMIC INTEGRITY AND STUDENTS' ETHICAL USE OF CHATGPT IN HIGHER EDUCATION > LITERATURE REVIEW
What the sources say
- Partly supportedcites Gozalo-Brizuela & Garrido-Merchan, 2023 sourceExperimental — accuracy not yet measured
The Conclusions and further work (p. 19) say generative AI models face bias from training data and that it needs to be controlled. The paper does not specifically report that ChatGPT or other language models reproduce stereotypes or narrow demographic representations, nor does it discuss governance mechanisms. (LLM, conf=0.84)
- Partly supportedcites Zhou et al., 2023 sourceExperimental — accuracy not yet measured
Section 4.5, “Data Cycle in IA Development” (page not provided), says AI bias can involve demographic features and data representativeness, and calls for measures to mitigate it. The paper does not show that language models such as ChatGPT reproduce stereotypes or narrow demographic representations; its discussion of ChatGPT elsewhere concerns other risks. (LLM, conf=0.78)
boostedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“While proper attribution allows students to benefit from AI without violating academic standards (Perkins, 2023), empirical studies indicate that failing to disclose AI assistance is commonly interpreted as academic dishonesty (Eke, 2023;Jarrah et al., 2023;Mansilla et al., 2022;Wardat et al., 2023).”
§ ACADEMIC INTEGRITY AND STUDENTS' ETHICAL USE OF CHATGPT IN HIGHER EDUCATION > LITERATURE REVIEW
What the sources say
- Not checked — no full text of the sourcecites Eke, 2023 source
- Partly supportedcites Jarrah et al., 2023 sourceExperimental — accuracy not yet measured
The “ChatGPT: Writing Tool or Plagiarism Without Attribution” section says that failing to acknowledge ChatGPT can lead to ethical issues and violations of academic integrity. But it describes concerns raised by a subset of articles, not empirical evidence that nondisclosure is commonly interpreted as dishonesty; the “Using ChatGPT in Academic Writing” section also says misconduct depends on institutional policies. Page numbers are unavailable in our copy. (LLM, conf=0.85)
- Not matchedcites Mansilla et al., 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Existing academic integrity frameworks (e.g., Borenstein & Howard, 2021;Eaton, 2022;Macfarlane et al., 2014) primarily address traditional forms of misconduct, such as plagiarism and cheating, focusing on moral reasoning and institutional codes rather than technology-mediated ethics.”
§ ACADEMIC INTEGRITY AND STUDENTS' ETHICAL USE OF CHATGPT IN HIGHER EDUCATION > RESEARCH MODEL
What the sources say
- Partly supportedcites Borenstein & Howard, 2021 sourceExperimental — accuracy not yet measured
Section 5 (AI ethics instruction; page unavailable) says that some ethics teaching focuses on professional codes and that courses should cultivate ethical reasoning. It does not assess academic integrity frameworks or say they primarily address plagiarism and cheating rather than technology-mediated ethics. (LLM, conf=0.83)
- Could not verifycites Eaton, 2022 sourceExperimental — accuracy not yet measured
The Introduction discusses plagiarism, cheating, and technologies used to detect misconduct; the Conclusions discuss ethical technology use. Neither section assesses existing academic integrity frameworks or says they prioritize moral reasoning and institutional codes over technology-mediated ethics. Searches for frameworks, moral reasoning, and policies found no passage establishing that comparison. (LLM, conf=0.72)
- Not checked — no full text of the sourcecites Macfarlane et al., 2014 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
89 more claims · 1 with their cited sources checked
Create a free account to see all 169 claims and every source verdict.
- Probably not a substantial claim. 24
- It may state two separate claims as one. 57
- The quote only partly states this claim. 1
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ ACADEMIC INTEGRITY AND STUDENTS' ETHICAL USE OF CHATGPT IN HIGHER EDUCATION > DATA ANALYSIS
“3.5 -4.0 4.10 4.00 3.70 3.30 4.00”
§ ACADEMIC INTEGRITY AND STUDENTS' ETHICAL USE OF CHATGPT IN HIGHER EDUCATION > DATA ANALYSIS
“Male 3.5 -4.0 4.15 4.05 3.80 3.40 4.10 Female Below 2.5 2.95 2.90 2.50 2.00 2.90 Female 2.5 -2.99 3.40 3.35 2.90 2.40 3.25 Female 3.0 -3.493.80 3.75 3.30 2.80 3.70 Female 3.5 -4.0 3.95 3.85 3.60 3.00 3.90”
Citations
Cites
The theory of planned behavior
Ajzen, I. (1991) · Organizational Behavior and Human Decision Processes · doi:10.1016/0749-5978(91)90020-t · Cited by 90,587
Cornally, N. (2015) · Springer Publishing Company eBooks · doi:10.1891/9780826164056.0005 · Cited by 36,743 · cited 2×
“…results confirm that students’ ethical orientations are central to shaping how generative AI is incorporated into academic work, consistent with the Theory of Planned Behavior (Ajzen, 1991) and Moral Intensity Theory (Jones, 1991), which link ethical awareness to responsible behavioral intentions.”
Pls-Sem: indeed a silver bullet
Hair, J. F., Ringle, C. M., & Sarstedt, M. (2011) · Journal of Marketing Theory and Practice · doi:10.2753/mtp1069-6679190202 · Cited by 21,631 · cited 1×
“PLS - SEM is robust to non - normal data distributions and performs well with moderate sample sizes (Hair et al., 2011, 2017; Sarstedt et al., 2017).”
Comparison of convenience sampling and purposive sampling
Etikan, I. (2016) · American Journal of Theoretical and Applied Statistics · doi:10.11648/j.ajtas.20160501.11 · Cited by 11,412 · cited 1×
“Consequently, sampling error may arise from underrepresentation of students with limited AI exposure or lower digital literacy, which may restrict the generalizability of findings beyond the sampled institutions and region (Etikan et al., 2015; Jager et al., 2017).”
The use of partial least squares path modeling in international marketing
Henseler, J., Ringle, C. M., & Sinkovics, R. R. (2009) · Advances in International Marketing · doi:10.1108/s1474-7979(2009)0000020014 · Cited by 10,926
ChatGPT for good? On opportunities and challenges of large language models for education
Kasneci, E., Sessler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier, E., et al. (2023) · Learning and Individual Differences · doi:10.1016/j.lindif.2023.102274 · Cited by 7,326 · cited 1×
“…programming, and problem - solving by leveraging deep learning techniques to generate contextually relevant responses (Atlas, 2023; Baidoo - Anu & Owusu Ansah, 2023; Else, 2023; Herft, 2023; Kasneci et al., 2023; Qadir, 2022; Sallam, 2023; Sok & Heng, 2023; Susnjak, 2022; Vieriu & Petrea, 2025).”
Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019) · International Journal of Educational Technology in Higher Education · doi:10.1186/s41239-019-0171-0 · Cited by 7,242 · cited 3×
“Furthermore, theoretical discussions often lack practical, evidence - based strategies to guide the ethical use of AI in educational contexts (Guerrero - Dib et al., 2020; Ramdani, 2018; Zawacki - Richter et al., 2019).”
The global landscape of AI ethics guidelines
Jobin, A., Ienca, M., & Vayena, E. (2019) · Nature Machine Intelligence · doi:10.1038/s42256-019-0088-2 · Cited by 6,220 · cited 1×
“Nevertheless, transparency remains in-consistently embedded in institutional policy frameworks, which are fragmented and still evolving (Jobin et al., 2019; Kazim & Koshiyama, 2021; Larsson & Heintz, 2020).”
Cohen, L., Manion, L., & Morrison, K. (2017) · doi:10.4324/9781315456539 · Cited by 6,018
Explanation in artificial intelligence: Insights from the social sciences
Miller, T. (2019) · Artificial Intelligence · doi:10.1016/j.artint.2018.07.007 · Cited by 5,302 · cited 1×
“Early governance literature conceptualizes transparency as a multidimensional and system - level construct rather than a narrow technical property (Hood, 2006; Larsson & Heintz, 2020; Margetts, 2011; Miller, 2019; Theodorou et al., 2017).”
Research works printed but not linked: 9
- Chin, W. W. (1998). The partial least squares approach to structural equation modeling. In G. A. Marcoulides (Ed.), Modern methods for business research (pp. 295-336). Lawrence Erlbaum. https://psycnet.apa.org/rec- ord/1998-07269-010
- Creswell, J. W., & Plano Clark, V. L. (2018). Designing and conducting mixed methods research approach (3 rd ed.). Sage.
- Hassan, Z. A., Schattner, P., & Mazza, D. (2006). Doing a pilot study: Why is it essential? Malaysian Family Physi- cian, 1(2-3), 70-73. https://pubmed.ncbi.nlm.nih.gov/27570591/
- Iqbal, U., & Iqbal, A. (2024). Assessing the effects of artificial intelligence on student cognitive skills: An inves- tigation into enhancement or deterioration. International Journal of Business, Analytics, and Technology, 2(1), 65- 75.
- Jeffrey, T. (2020). Understanding college student perceptions of artificial intelligence. Systemics, Cybernetics and Informatics, 18(2), 8-13. https://iiisci.org/journal/pdv/sci/pdfs/HB785NN20.pdf
- Luckin, R., Holmes, W., Griffiths, M., & Forcier, L. B. (2016). Intelligence unleashed: An argument for AI in education. Pearson. https://static.googleusercontent.com/media/edu.google.com/en//pdfs/Intelligence-Unleashed- Publication.pdf
- Mansilla, E. B. R., Castillo-Acobo, R. Y., Puma, E. G. M., Maquera, Y. M., Gonzales, J. L. A., & Vasquez- Pauca, M. J. (2022). Stress in university teachers in the framework of the post-pandemic face-to-face aca- demic resumption. Journal of Medicinal and Chemical Sciences, 5(6) 1040-1047. https://doi.org/10.26655/JMCHEMSCI.2022.6.17 · doi:10.26655/jmchemsci.2022.6.17
- Mihaella, A., & Petrea, G. (2025). The impact of artificial intelligence on students' academic performance devel- opment. Education Sciences, 15(3), 343.
- Porayska-Pomsta, K., & Luckin, R. (2022). Issues with data, bias, and fairness. In C. de la Higuera & J. Iyer (Eds.), AI for teachers: An open textbook. Press Books. https://pressbooks.pub/aiforteachers/chapter/issues- with-data-bias-and-fairness/
Cited by
Benchmarking AI-Generated Text Detection Tools in Academic Writing: An Experimental Evaluation
Alaqad, M. H., Alkindi, G., Hussin, M., Sapar, A. A., & Alaggad, L. A. (2026) · International Congress on Human-Computer Interaction, Optimization and Robotic Applications (ICHORA) · doi:10.1109/ichora69329.2026.11537020
“Detection tools primarily exploit the observation that LLM-generated text tends to exhibit lower perplexity under the generating model and lower “burstiness” less variation in se ntence complexity than human-written text [11].”
InfluentialMárquez, G. B., Capuñay, J. E. C., Calcina, J. G. M., Rosales, L. L. R., Atahualpa, R. L. C., & Sebastián, M. G. M. (2026) · Prohominum · doi:10.47606/acven/ph0529
Predictors of the ethical use of generative artificial intelligence in higher education
Cruz, Y. B. T., Sánchez, H. A. C., Pita, I. G. A., Rodríguez-Álava, L. A., Carreño-Mendoza, Á. L., Cedeño-Barreto, M., Castillo, G., Villavicencio, J. C., & Moreira-Chóez, J. S. (2026) · Frontiers in Education · doi:10.3389/feduc.2026.1942426 · Cited by 1
Mahmood, K. A., & Mustafa, Z. (2026) · Academic Journal of International University of Erbil · doi:10.63841/iue33564
Metadata
- Authors
- Shishakly, R., & Nachouki, M.
- Affiliations
- Ajman University (AE)
- Venue
- Journal of Information Technology Education: Research
- Publisher
- Informing Science Institute
- Published
- 2026
- Volume
- 25
- Pages
- 08
- ISSN
- 1539-3585 (ISSN-L)
- 1547-9714
- OpenAlex
- W7135418306
- Cited by
- 15
- References
- 147
- Citations by year
- 2026: 15
- FWCI
- 76.55
- Type
- journal article
- Language
- en
- Open access
- diamond
- doi.org
- Licence
- Locations
- Journal of Information Technology Education Research (publishedVersion, cc-by-nc)
- Topics
- Social Sciences › Education › Academic integrity and plagiarism
- Social Sciences › Education › Artificial Intelligence in Education
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Keywords
- academic integrity, ChatGPT, generative AI, ethical use of AI, trust in AI, perceived usefulness, digital literacy, academic performance, gender differences, disciplinary differences, plagiarism avoidance, bias awareness, responsible AI use, AI ethics training, Gulf universities
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-05, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-05 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.