Interplay of rationality and morality in using ChatGPT for academic misconduct

ChatGPT’s capability to generate original content on any topic represents a concern for academic institutions, since students can plagiarise the AI-created content with a low probability of being caught. While prior research on academic misconduct emphasized the combined effects of perceived costs and benefits in explaining students’ decisions to cheat, the low detectability of ChatGPT-based plagiarism significantly reduces the cost of being caught, making rational choice theory less powerful in explaining academic misconduct in the AI era. Considering the effects of both rational decision-making and moral considerations, this study shows that individuals experiencing moral disengagement are more likely to engage in ChatGPT-based plagiarism. Other significant predictors include perceived benefits, punishment severity, and informal sanctions. Furthermore, moral disengagement enhances the effects of formal sanctions on ChatGPT-based plagiarism. The results have implications for professors, academic administrators, and plagiarism detection software designers to enforce academic integrity in the ChatGPT era.