Generative AI (GenAI) is increasingly embedded in higher education, yet evidence on its implications for students’ academic engagement and psychological experiences remains mixed. One possible reason is that prior research has often focused on how much students use AI and their general confidence in task completion, while paying less attention to how they use AI and how they attribute AI-supported achievement. To address this gap, this study distinguishes reflective from thoughtless AI use, examines academic impostor syndrome as a self-evaluative mechanism linking AI use styles to academic engagement, and tests perceived AI policy clarity as a contextual moderator. A two-wave survey of 478 Chinese university students showed that reflective AI use was negatively associated with academic impostor syndrome, whereas thoughtless AI use showed the opposite pattern. Academic impostor syndrome, in turn, was negatively associated with engagement and mediated both pathways. Perceived AI policy clarity amplified these patterns. These findings suggest that GenAI integration should be understood not only as a question of adoption or efficiency, but also of interaction quality and competence attribution. The study highlights the importance of cultivating reflective AI literacy and developing institutional policies that are clear yet psychologically attuned to students’ self-evaluative concerns.
Wang et al. (Wed,) studied this question.