When Justification Meets Expectation and Control Fuels Detachment: Paradoxes of AI-Assisted Plagiarism

The rapid adoption of generative AI in higher education has introduced new tensions in academic integrity and ethical decision-making. This study examines AI-assisted plagiarism among 310 undergraduate and graduate students at an international European university by integrating moral disengagement theory with the indirect misbehaviors framework. Using Smart PLS-SEM and PROCESS analyses, the findings reveal two paradoxes that shape students’ intentions and cognitive outcomes. The first is the Justification-Expectation Paradox, where moral disengagement increases plagiarism intentions, but this effect becomes weaker when subjective norms are strong. Students may justify using AI tools to complete academic work, yet these justifications lose their influence when social expectations emphasize that such behavior is unacceptable. Shame does not directly affect intention, indicating that normative pressures outweigh emotional discomfort. The second is the Control-Detachment Paradox, where plagiarism intentions reduce cognitive engagement and increase psychological detachment, while perceived behavioral control alters these relationships. Students with high control remain more engaged and less detached, whereas those with low control show stronger withdrawal from authentic learning. These findings extend the indirect misbehaviors framework to generative AI contexts and highlight how ethical reasoning, social expectations, and perceived control interact in shaping AI-related misconduct. The study emphasizes the need to strengthen subjective norms and promote responsible AI literacy, since traditional deterrence and rational cost-benefit mechanisms played limited roles.