The DOI matches its Crossref record; 80 of the 82 references it cites were found in a registry or an index; 10 of the 35 sources its claims cite were checked for support.
Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students
Open accessAbstract
Academic dishonesty remains a persistent concern for educational institutions, threatening the reputation of universities. The emergence of Artificial Intelligence (AI) tools exacerbates this challenge as they can be used for chatting but also for cheating. Several scientific papers have analyzed the advantages and risks of using AI tools like ChatGPT in academia. On the one side it offers increased accessibility of information and facilitates personalized learning. On the other side several risks are associated with the usage of ChatGPT, including the generation of inaccurate information, and the facilitation of academic dishonesty like cheating and plagiarism. This study examines how students’ Perception of Academic Dishonesty (PAD) and their perception of the risks associated with using ChatGPT, are linked to a reduced use frequency and intention to use this AI tool. A sample of 468 undergraduate students answered an online survey. Our findings indicate a negative and significant relationship between perceived risk, and both use frequency and intention to use ChatGPT in the future. Additionally, perceived risk mediates the relationship between PAD and ChatGPT usage patterns. This is the first study to demonstrate that students with high PAD present higher levels of awareness of the risks associated with using ChatGPT, which is associated with lower frequency of use and intention to use this AI tool in higher education. These results highlight the importance of integrating such technologies into the university, while carefully considering ethical issues and students’ perceptions of risk.
Overview
CoChat Summary
TL;DR
In a sample of 468 Spanish undergraduates who had used ChatGPT, perceived risk was negatively related to both use frequency and intention to use, and it mediated the relationship between Perceived Academic Dishonesty (PAD) and both outcomes. PAD was not directly related to use frequency or intention to use.
Key findings
- Perceived risk was significantly and negatively associated with use frequency (β = −0.20, p < .001) and intention to use ChatGPT (β = −0.46, p < .001), supporting H1a and H1b.
- PAD was not directly related to use frequency (β = 0.04, p = .22) or intention to use (β = −0.05, p = .33), rejecting H2a and H2b.
- PAD was positively related to perceived risk (ρ = 0.13, p < .05), and perceived risk mediated PAD’s relationship with use frequency (95% CI [−0.05, −0.01]) and intention to use (95% CI [−0.11, −0.02]), supporting H3a and H3b; the authors describe this as indirect-only mediation.
- Perceived risk correlated negatively with use frequency (ρ = −0.21, p < .001) and intention to use (ρ = −0.33, p < .001); use frequency and intention to use were positively correlated (ρ = 0.52, p < .001).
Methods
- Cross-sectional online survey distributed via Google Forms at the University of the Balearic Islands in October–November 2023; of 737 respondents, 468 who reported having used ChatGPT were analyzed. Participants averaged 20 years old (SD = 4.25); 55% were female, 44% male, and 1% non-binary.
- Measures included ChatGPT use frequency, intention to use (3 items; Cronbach’s α = 0.98, McDonald’s ω = 0.98), perceived risk (5 items plus one added item; α = 0.78, ω = 0.79), and PAD (15 items; α = 0.96, ω = 0.96).
- Analyses used Jamovi 2.3.18; Spearman correlations were used because PAD was non-normal, and mediation was tested with GLM Mediation using 10,000 bootstrap samples and bias-corrected confidence intervals.
Limitations
- The paper states limitations including a non-probability convenience sample from a single university, lack of an a priori power analysis, cross-sectional design limiting causal interpretation, PAD’s non-normal distribution requiring non-parametric methods, and possible social desirability bias.
- The paper also notes that it captures general risk perceptions rather than specific dishonest ChatGPT behaviors, and calls for future research on specific academic-integrity violations and longitudinal designs.
- Evident from the text: the sample includes only students who had used ChatGPT, and the reported AVE values for PAD (0.46) and perceived risk (0.40) fall slightly below the commonly recommended 0.50 threshold, although the authors argue composite reliability supports the constructs.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
9 Valid · 0 Not valid · 13 Not checked · 33 Own results and interpretations
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“Prior research in social psychology and technology adoption has shown that higher perceived risk decreases the likelihood of engaging in a given behavior (Chiu et al., 2023;Essel, 2022;Lee et al., 2024;Mvondo et al., 2023;Sultana, 2020).”
§ Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students > 2.2 Perceived risk and ChatGPT usage
What the sources say
- Not checked — no full text of the sourcecites Chiu et al., 2023 source
- Not matchedcites Essel, 2022
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Supportedcites Lee et al., 2024 source
The Relative risk vs. benefit perception section says people are less likely to adopt new technologies when they perceive high risks. The Additional analyses section reports that respondents who believed ChatGPT’s risks outweighed its benefits expressed less willingness to use it. Together, these support the cited claim in the technology-adoption context. (LLM, conf=0.88)
plainstates a cause- 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.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“In contrast, students who score high on the honestyhumility dimension-a personality trait closely associated with ethical behaviortend to avoid using AI for dishonest purposes (Greitemeyer & Kastenmüller, 2023).”
§ Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students > 2.3 Perceptions of academic dishonesty and ChatGPT usage
What the sources say
- Partly supportedcites Greitemeyer & Kastenmüller, 2023 sourceExperimental — accuracy not yet measured
The Results report that higher Honesty-Humility was associated with lower intention to use chatbot-generated texts for academic cheating. However, the Discussion’s “Limitations and future research” section says the study measured intentions, not actual use. The claim’s wording about students avoiding AI use therefore goes beyond what the study established. (LLM, conf=0.87)
plainstates an association- 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
“Bergström et al. (2024), who found that Finnish university students were reluctant to use ChatGPT in academic settings when they perceived it as ethically problematic, particularly due to concerns about its potential misuse for cheating.”
§ Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students > 2.3 Perceptions of academic dishonesty and ChatGPT usage
What the sources say
- Partly supportedcites Bergström et al., 2024 sourceExperimental — accuracy not yet measured
The Research method section describes a survey of 104 business-school students in Finland, and section 5.2 reports that students judged cheating with ChatGPT wrong, particularly during tests. But the study asked how likely respondents thought their classmates were to engage in these behaviors, not whether respondents themselves were reluctant to use ChatGPT. It also found some academic uses acceptable. Page numbers are unavailable in our copy. (LLM, conf=0.88)
plainstates 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
“our results are consistent with previous research that showed a link between perceived risk and the intention to use ChatGPT”
§ Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students > 5 Discussion
What the sources say
- Not matchedcites Gonzalo & Enriquez, 2023
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
- Partly supportedcites Sallam et al., 2023 sourceExperimental — accuracy not yet measured
The Discussion (Principal Results; page unavailable) says risk perception significantly influenced Jordanian health care students’ attitudes and usage of ChatGPT. That supports a link with adoption-related behavior, but the paper reports attitudes and use rather than a measured link with intention to use ChatGPT. (LLM, conf=0.82)
plainstates an association- 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.
- Not matched
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Additionally, perceived risk mediates the relationship between PAD and ChatGPT usage patterns.”
- Also stated in: § 4.1 Hypotheses testing · plain
“Third, PAD was indirectly associated with both use frequency (95% CI [-0.05, -0.01]) and intention to use (95% CI [-0.11, -0.02]) through perceived risk.”
- Also stated in: § 4.1 Hypotheses testing · boosted
“Since the confidence intervals did not include zero, we conclude that perceived risk medi-ates the relationship between PAD and both outcome variables.Therefore, H3a and H3b were supported.”
§ Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students > Abstract
boostedstates a cause- 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.
- Also stated in: § 4.1 Hypotheses testing · plain
28 more claims
Create a free account to see all 55 claims and to verify the 1 reference still open.
- Probably not a substantial claim. 14
- It may state two separate claims as one. 10
- The quote only partly states this claim. 2
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Recent work therefore argues that the educational value of generative AI depends less on access alone than on whether students use it to construct and critique knowledge rather than outsource reasoning, and on whether students remain sufficiently risk-aware to avoid academically dishonest or unreflective use (Pallant et al., 2025; Costa and Murphy, 2025; Ortiz-Bonnin and Blahopoulou, 2025).”
From College students' metacognitive awareness of generative-AI reliance: an experimental study of decision confidence and attribution bias (2026) · cites this paper as Ortiz-Bonnin and Blahopoulou, 2025
Experimental — accuracy not yet measured
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
§ Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students > 2.4 Perception of academic dishonesty, perceived risk, and ChatGPT usage
“Thus, our third hypotheses are: H3a: Perceived risk will act as a partial mediator between PAD and use frequency of ChatGPT. H3b: Perceived risk will act as a partial mediator between PAD and intention to use ChatGPT.”
§ Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students > 4.1 Hypotheses testing
“According to Meule (2019) we found an indirect-only mediation because indirect effects are significant whereas the direct effect (from independent variable to dependent variable) and total effects are not significant.”
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
§ Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students > 3.4 Data analysis
“We employed a non-parametric resampling technique known as bootstrapping to derive confidence intervals based on percentiles (MacKinnon et al., 2004).”
§ Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students > 3.4 Data analysis
“The indirect effect was computed using 10,000 bootstrap samples to generate Bias-Corrected (BC) confidence intervals (CI).”
Citations
Cites
Evaluating Structural Equation Models with Unobservable Variables and Measurement Error
Fornell, C., & Larcker, D. F. (1981) · Journal of Marketing Research · doi:10.2307/3151312 · Cited by 109,314
Evaluating structural equation models with unobservable variables and measurement error
Fornell, C., & Larcker, D. F. (1981) · Journal of Marketing Research · doi:10.1177/002224378101800104 · Cited by 109,190 · cited 1×
“According to Fornell and Larcker (1981), AVE values below 0.50 can be considered acceptable when composite reliability exceeds 0.60.”
InfluentialBaron, R. M., & Kenny, D. A. (1986) · Journal of Personality and Social Psychology · doi:10.1037/0022-3514.51.6.1173 · Cited by 91,532
A new criterion for assessing discriminant validity in variance-based structural equation modeling
Henseler, J., Ringle, C. M., & Sarstedt, M. (2015) · Journal of the Academy of Marketing Science · doi:10.1007/s11747-014-0403-8 · Cited by 38,459 · cited 1×
“Discriminant validity was evaluated using the Heterotrait-Monotrait (HTMT) ratio of correlations, following Henseler et al. (2015).”
When to Use and How to Report the Results of PLS-SEM
Hair, J. F., Risher, J. J., Sarstedt, M., & Ringle, C. M. (2019) · European Business Review · doi:10.1108/ebr-11-2018-0203 · Cited by 27,541 · cited 4×
“The VIF values were 1.01, which is below the commonly recommended threshold of 5 (Hair et al., 2019).”
Influential(2014) · Climate Change 2013 – The Physical Science Basis · doi:10.1017/cbo9781107415324.004 · Cited by 23,880
A BEHAVIORAL MODEL OF RATIONAL CHOICE
Simon, H. A. (1955) · The Quarterly Journal of Economics · doi:10.2307/1884852 · Cited by 15,877
Crime and Punishment: An Economic Approach
Becker, G. S. (1968) · Journal of Political Economy · doi:10.1086/259394 · Cited by 13,999
Confidence Limits for the Indirect Effect: Distribution of the Product and Resampling Methods
MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004) · Multivariate Behavioral Research · doi:10.1207/s15327906mbr3901_4 · Cited by 7,806
Sallam, M. (2023) · Healthcare · doi:10.3390/healthcare11060887 · Cited by 3,126 · cited 1×
“Drawing on a substantial sample size, this research offers a broader and more representative understanding of the phenomenon (Lee et al., 2024 ; Sallam, 2023 ).”
Research works printed but not linked: 2
- Elias, R. Z. (2015). The effects of machiavellianism and cultural dimensions on business students’ general distrust of corporations. Journal of Business Studies Quarterly, 6(4), 21. https://www.proquest.com/scholarly-journals/effects-machiavellianism-cultural-dimensions-on/docview/1694649856/se-2
- Kahneman, D. (2011). Fast and slow thinking. Allen Lane and Penguin Books.
Cited by
Wu, C., & Carroll, J. M. (2025) · TechTrends · doi:10.1007/s11528-025-01150-5 · Cited by 16
Drivers of academic dishonesty: situational pressures versus student motivation
Yin, X., Lai, Q., & Guo, Y. (2025) · International Journal of Adolescence and Youth · doi:10.1080/02673843.2025.2560649 · Cited by 8
“As Bonnin & Blahopoulou (2025) argue, perceived academic dishonesty constitutes a perceptual construct reflecting students' judgments about the acceptability of deceptive academic behaviours.”
Pedraja-Rejas, L., Vega, P. L., & Huanca, P. R. (2025) · Behavioral Sciences · doi:10.3390/bs16010011 · Cited by 7
“This well-founded scepticism, coupled with the lack of clear regulations which, according to Ortiz-Bonnin and Blahopoulou (2025), generates confusion and fear of being judged, acts as a significant barrier to authentic and sustained engagement.”
Wang, Y., Gu, X., & Li, S. (2026) · Humanities and Social Sciences Communications · doi:10.1057/s41599-026-06555-y · Cited by 2
Huang, Y., Zhang, K., & Pan, R. (2026) · Higher Education · doi:10.1007/s10734-026-01630-4 · Cited by 3
Metadata
- Authors
- Ortiz-Bonnin, S., & Blahopoulou, J.
- Affiliations
- Universitat de les Illes Balears (ES)
- Corresponding author
- Blahopoulou, Joanna
- Venue
- Social Psychology of Education
- Published
- 2025-05-28
- Volume
- 28
- Issue
- 1
- ISSN
- 1381-2890 (ISSN-L)
- 1573-1928
- OpenAlex
- W4410841095
- Cited by
- 80
- References
- 72
- Citations by year
- 2026: 70 · 2025: 10
- FWCI
- 66.53
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- Social Psychology of Education (publishedVersion, cc-by)
- Funders
- Universitat de Les Illes Balears
- Topics
- Social Sciences › Education › Academic integrity and plagiarism
- Social Sciences › Education › Artificial Intelligence in Education
- Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
- Keywords
- academic dishonesty, ChatGPT, generative AI in education, risk perception, intention to use AI, usage frequency, plagiarism, cheating, undergraduate students, AI ethics
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students. https://app.cochat.ai/papers/10.1007/s11218-025-10080-2
- The paper
- Ortiz-Bonnin, S., & Blahopoulou, J. (2025). Chat or cheat? Academic dishonesty, risk perceptions, and ChatGPT usage in higher education students. Social Psychology of Education. https://doi.org/10.1007/s11218-025-10080-2
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.