This study investigates the factors influencing accounting students' misuse of ChatGPT for academic dishonesty, drawing on the Fraud Triangle Theory (opportunity, pressure, and rationalization). It also examines the role of institutional policies and the moderating effect of academic support in shaping students' ethical decisions.
A quantitative design was employed using survey data from 474 accounting students in Bangladesh who had experience using ChatGPT. Structural Equation Modeling (PLS-SEM) was applied to test the hypothesized relationships among Opportunity (OPP), Pressure (PRE), Rationalization (RAT), Institutional Policies (POL), and Academic Dishonesty (DIS), while assessing Academic Support (AS) as a moderating variable.
Results reveal that opportunity and pressure significantly increased students' likelihood of engaging in academic dishonesty, while rationalization and institutional policies showed insignificant effects. The results reveal that academic support significantly moderates the relationships between Pressure, rationalization, and academic dishonesty, indicating that these effects are more potent at higher levels of academic support. However, it did not significantly moderate the relationship between opportunity and academic dishonesty. The model explained 43% of the variance in academic dishonesty, highlighting the explanatory strength of the framework.
The findings emphasize the need for universities to strengthen academic integrity policies, integrate AI ethics education, and expand academic support systems, such as tutoring, counseling, and mentoring. These measures can reduce students' reliance on ChatGPT for dishonest purposes and foster a culture of academic integrity.
This study is among the first to apply the Fraud Triangle Theory to AI-assisted academic dishonesty, introducing academic support as a novel moderator. It offers theoretical insights and practical guidance for educators and policymakers in emerging economies.
