Mechanisms of academic stress affecting AI-assisted cheating behaviour in college students: a mixed methods study

The emergence of Generative Artificial Intelligence (GAI) has provided college students with tools to mitigate academic stress but has also facilitated AI-assisted cheating. However, existing research offers limited insight into how academic stress influences cheating behaviours and the underlying psychological mechanisms, leaving a significant gap in understanding. This study aims to develop and validate a theoretical model explaining how academic stress influences college students’ intentions to engage in AI-assisted cheating. Grounded in the Theory of Planned Behaviour (TPB), this research employed a mixed-methods approach. Quantitative data were analysed to explore relationships among academic stress, academic self-efficacy, attitudes toward AI-assisted cheating, and cheating intentions. A survey analyzed 458 valid questionnaires from students, while follow-up qualitative interviews with 21 students who admitted to using AI to assist in cheating provided in-depth analysis. Findings indicate that academic stress directly influences students’ AI-assisted cheating intentions and indirectly does so through academic self-efficacy and attitudes as mediators. Qualitative interviews reinforced these findings by offering contextual explanations. This study reveals both direct and indirect pathways linking academic stress to AI-assisted cheating, contributing to a deeper theoretical and practical understanding of academic misconduct in the GAI era.