A Criminological Mixed-Methods Analysis of University Policies and Strategies Against AI Cheating: A Case Study of Novel Deviance and Cultural Lag

University policies regarding AI-facilitated cheating are premised on classic criminological notions of rational choice positing that individuals will refrain from actions, such as academic dishonesty, because of the risks of being caught and punished. In the present study, we examine the efficacy of such policies and the applicability of Rational Choice Theory (RCT) to AI-facilitated cheating with a large scale, mixed-methods experiment ( N = 3,298 college students). In support of RCT, we find that students’ perceived likelihood of getting caught and perceived severity of punishment significantly reduced their likelihood of using AI to cheat. However, we do not find that the tested university practices and policies influence students’ perceptions of these risks. Our analysis suggests that this is due to the novelty of AI. First, there is a “cultural lag” among AI use and its regulation in society. As such, students’ decisions to use AI to cheat are primarily driven by various personal factors, rather than the situational elements such as school policy. These include their own moral assessments of cheating and their personal definitions of what constitutes cheating with AI. Secondly, as using AI to cheat is a novel form of deviance, there is no consensus on what constitutes a high (low) risk situation. We discuss implications for RCT more broadly and make suggestions for university administrators and instructors for addressing this growing form of academic misconduct.