AI’s learning paradox: how business students’ engagement with AI amplifies moral disengagement-driven misconduct

Abstract

Artificial intelligence (AI) in higher education creates a learning paradox, enhancing productivity while enabling hard-to-detect misconduct, challenging ethical boundaries, and university policies. Drawing on moral disengagement (MD) theory, this study examines how AI engagement conditions, captured by the Motive, Means, Opportunity (MMO) framework, amplify MD’s effect on misconduct among graduate business students. Self-Regulated Learning (SRL) offers a learning process lens to locate where MD and its MMO conditions unfold within the learning cycle. Survey data from 226 UK-based students shows that MD predicts AI misconduct, with amplification from AI-related factors (usefulness, habit, obsessive passion, prompt engineering skill) and past misconduct. Policy enforcement mitigates this effect, while policy clarity is effective only when paired with enforcement. Unexpectedly, high-performing students are more likely to act on MD when scanning for misconduct opportunities. Our findings underline how AI engagement undermines ethical regulation, offering insights for institutional policy in AI-enabled learning environments.