Beyond Didactics and Instructing: Adapting Team-Based Learning and Reflective Practice to Develop Ethical AI Engagement in Medical Students

Background: Medical students widely adopt generative AI tools yet lack AI literacy, awareness, and technical competence required for critical evaluation and ethical use. AI ethics education remains largely absent from medical curricula.

Methods: A quasi-experimental pre-post design was used within a required Medical Humanities and Wellness course at a Doctor of Osteopathic Medicine (DO) programme. A 2-hour session combined Team-Based Learning (TBL) with ERA (Experience-Reflection-Action) Reflective Practice. Pre-session (n = 113) and post-session (n = 104) surveys measured five domains: awareness, perception, knowledge, confidence, and attitude. Fifteen collaborative group reflections provided qualitative data analysed thematically within the ERA framework.

Results: All five domains improved significantly. AI understanding rose markedly (U = 3,134.5, r = 0.462, p <  0.001); DAPI (Data, Algorithm, Platform, Infrastructure) knowledge showed the largest effect (r = 0.642, p <  0.001); confidence in critically evaluating AI outputs increased from 19.5% to 92.3%; and support for formal AI curriculum integration nearly doubled (30.1% to 59.6%; χ²(2) = 18.83, p <  0.001). AI-related anxiety declined significantly (p = 0.030). Qualitative analysis yielded nine ERA-structured themes converging on responsible use, recalibrated expectations, and ethical stewardship.

Conclusions: A single TBL and Reflective Practice session produced meaningful improvements in AI literacy and ethical engagement. The ERA framework effectively translated knowledge into behavioral commitments. These findings offer an evidence-based, replicable model for AI ethics integration in medical education. This work also inspires educators to seek practical ways of integrating AI literacy and AI ethics in medical school curricula.