An Interactive, Project-Based AI in Healthcare Curriculum for Clinicians and Medical Educators

Abstract

Introduction A persistent skills gap separates the rapid development of artificial intelligence (AI) tools in medicine from clinicians’ ability to critically evaluate, adopt, and govern them. Existing AI-in-healthcare education is typically delivered through passive, lecture-based formats that do not require learners to apply frameworks to real clinical or institutional problems. Methods We designed and piloted a 10-week, interactive AI in Healthcare curriculum using a Socratic teaching method: each concept is taught, followed by a targeted question the learner must answer before advancing; every task submission receives detailed individualized correction. The curriculum requires each learner to develop an original AI project addressing a real clinical or educational problem, culminating in a published GitHub repository documenting a structured readiness-objectives-adoption-data (R.O.A.D.) analysis, algorithm selection, an ethics and governance framework, and a validation plan. Results Pilot delivery of the curriculum’s foundational module demonstrated feasibility of the interactive correction model, with measurable improvement across four graded learner tasks, including increased precision in formulating measurable, evidence-based project objectives following targeted instructor correction. Discussion This curriculum offers a generalizable, no-cost model for closing the AI evaluation skills gap in healthcare education, combining interactive correction with a mandatory, learner-defined capstone project. We discuss plans for formal assessment validation and full-cohort implementation.