Ethical Judgments and Weakness of Will in the Misuse of Generative AI: The Case of Academic Essay Production among Higher Education Students

Abstract Generative artificial intelligence (GenAI) use in university presents both benefits and risks for academic integrity. Whereas its potential to enhance writing, support study practices, and stimulate creativity is well recognized, concerns are growing about its misuse in producing academic essays. This study examines the issue from two angles: the ethical judgments guiding students’ decision-making and the transition from legitimate to fraudulent uses. We draw on weakness-of-will accounts as a theoretical lens to interpret how routinized legitimate GenAI use may facilitate misuse intentions. This transition can be framed through behavioral science concepts like moral hazard or the law of the hammer. An explanatory framework was assessed using partial least squares–structural equation modeling. Fieldwork involved 151 students of social sciences programs from two Spanish higher education institutions. Three ethical dimensions from a Multidimensional Ethics Scale (moral justice, relativism, contractualism) were measured, together with a formative construct capturing legitimate uses of GenAI (linguistic editing of essays and study support), plus sociodemographic controls. Results show that the model explains nearly 70% of the variance in intentions to misuse GenAI. Moral justice and relativism emerge as significant predictors, whereas contractualism shows no relevant effects. Importantly, legitimate uses of GenAI increase the likelihood of misuse, acting as a facilitator. Thus, misuse of GenAI cannot be reduced to individual dishonesty but is better understood as a complex phenomenon shaped by ethical frameworks and by temptation mechanisms associated with routinized licit use. These findings highlight the need for universities and educators to reconsider pedagogical strategies and integrity policies in response to GenAI’s growing role in education.