Academic self-perceptions and AI use among health professions students: the serial mediating roles of learned helplessness and academic-clinical learning loneliness.

Objective: This study inspected the direct and indirect relationships between academic self-perceptions and artificial intelligence (AI) use among health professions students. More specifically, it investigated if learned helplessness and academic and clinical learning loneliness independently and sequentially mediated the relationships of academic wellbeing, academic incompetence, impostor syndrome and overall academic self-perceptions with AI use. Methods: A quantitative, cross-sectional, and correlational design was employed. The sample consisted of 650 undergraduate health professions students aged 18-25 years. Data were collected employing the Academic Self-Perceptions Scale, Learned Helplessness Scale, Academic and Clinical Learning Loneliness Scale, and AI Use Scale. Descriptive statistics, Pearson correlations and serial mediation analyses were conducted using Hayes' PROCESS Model 6. A total of four diverse models were tested, with academic wellbeing, academic incompetence, impostor syndrome and overall academic self-perceptions entered as predictors. Indirect effects were estimated using 5,000 bootstrap samples and 95% bias-corrected confidence intervals. Results: Academic wellbeing and positive overall academic self-perceptions were negatively associated with learned helplessness, academic and clinical learning loneliness, and AI use. In contrast, academic incompetence and impostor syndrome were positively linked to all three variables. Learned helplessness positively predicted academic and clinical learning loneliness, and both mediators were positively ascribed to AI use. All specific indirect effects through learned helplessness and academic and clinical learning loneliness were significant. The serial indirect effects through learned helplessness followed by academic and clinical learning loneliness were also significant in all four models, indicating partial mediation. Discussion: The findings spotlight that AI use of health professions students is related not only to technological accessibility but also to students' academic self-evaluations, perceived control over learning, and connection with academic and clinical learning environments. Positive academic self-perceptions can reduce AI use by limiting helplessness and learning-related loneliness, whereas academic incompetence and impostor syndrome may increase AI use through the opposite process. The results highlight the significance of supporting academic competence, learner agency, relational connection, and responsible AI use in health professions education.