The psychological mechanisms and behavioral determinants of academic integrity in the age of artificial intelligence.

Objective: This study aimed to examine the psychological and behavioral determinants of AI-assisted academic dishonesty among university students through an integrated model. Specifically, the study investigated whether academic procrastination, learned helplessness, and academic self-efficacy predict cheating tendency; whether cheating tendency predicts AI-assisted academic dishonesty; whether AI use moderates the relationship between cheating tendency and academic dishonesty; and whether social and contextual factors significantly predict AI-assisted academic dishonesty. Methods: The study employed a quantitative, cross-sectional, correlational survey design. A total of 1,045 undergraduate students from different academic disciplines participated voluntarily in the study. Data were collected using seven measurement instruments and a personal information form. Descriptive statistics, Pearson correlation analyses, multiple regression analyses, and Hayes' PROCESS Model 14 were used to test the proposed moderated mediation model. Bootstrap resampling with 5,000 samples was applied to estimate indirect effects and 95% confidence intervals. Results: The findings showed that academic procrastination and learned helplessness positively predicted cheating tendency, whereas academic self-efficacy negatively predicted it. Cheating tendency significantly predicted AI-assisted academic dishonesty, and the interaction term indicated that the association between cheating tendency and AI-assisted academic dishonesty was stronger at higher levels of AI use. Conditional indirect effect analyses further demonstrated that cheating tendency mediated the effects of academic procrastination, learned helplessness, and academic self-efficacy on AI-assisted academic dishonesty, and these indirect effects became stronger at higher levels of AI use. In addition, social norms, peer behaviors, family attitudes, insufficient sanctions, teacher attitude, high expectations, and adverse conditions significantly predicted AI-assisted academic dishonesty, whereas ethical and moral education emerged as a negative predictor. Discussion: The findings indicate that AI-assisted academic dishonesty should be understood as a multilevel outcome shaped by the interaction of psychological vulnerabilities, cognitive tendencies, technological affordances, and socio-contextual influences. The study contributes to the academic integrity literature by showing that AI use does not merely accompany dishonest tendencies but amplifies their translation into behavior. These results highlight the need for psychologically informed, ethically grounded, and institutionally supported interventions to reduce academic dishonesty in AI-enhanced higher education environments.