Eastwood K.W.Allali D.Leela-Amornsin S.Desbiens J.P.Szulewski A.Mahidol University2026-02-282026-02-282026-01-01Canadian Journal of Emergency Medicine (2026)14818035https://repository.li.mahidol.ac.th/handle/123456789/115428This work describes the use of participatory action research to develop an artificial intelligence (AI)-augmented, peer-driven, case-based, and simulation-enhanced framework for senior emergency medicine trainees. It has been applied to enhance knowledge acquisition for small-group self-directed study in resuscitation medicine. Trainees engaged in structured learning cycles over 6 months, based on the principles of ‘desirable-difficulty’ and deliberate-practice. It incorporated peer-selected pre-reading, case-based discussions, high-fidelity simulations, and spaced-repetition flashcard review. A key innovation is the use of generative AI tools to supplement these activities, and follow evidence-based prompt engineering. The participants refined self-study methods through iterative evaluation. AI-generated questions facilitated retrieval-based learning, and flashcard integration enhanced knowledge retention. Simulation-based reinforcement contributed to the ‘desirable-difficulty’ through the clinical application of learned concepts. Self-reported recall improved over time. This structured, self-directed approach supports effective learning in resuscitation medicine. AI and peer-driven strategies augment knowledge retention. This methodology offers adaptability for broader medical education settings.MedicineArticleSCOPUS10.1007/s43678-026-01118-12-s2.0-1050306646441481804341706282