Generative AI for clinical reasoning: A scoping review

Objectives: To explore how generative artificial intelligence (AI) supports clinical reasoning development through simulation-based teaching in undergraduate health professions education. Design: Scoping review. Data Sources: CINAHL, ERIC, PubMed, ScienceDirect, and Web of Science databases. Review Methods: A systematic search was conducted to identify studies exploring the integration of generative AI in simulation-based learning. Inclusion criteria focused on undergraduate health professions education and clinical reasoning outcomes. Results: Six studies with a total of 492 participants met the inclusion criteria. Generative AI was used to create simulation scenarios, virtual patients, provide feedback, analyze student performance, and support inquiry-based learning. Four studies reported significantly improved clinical reasoning outcomes with AI-assisted teaching. One study reported the comparability between AI-generated feedback and expert feedback, though expert input remained superior in complex cases. Conclusions: The integration of generative AI into simulation-based education is in its early stages. Most studies lacked theoretical frameworks and used diverse outcome measures, limiting comparability and generalizability. Future research should adopt theory-driven designs and standardized assessment tools to better evaluate the impact of generative AI on clinical reasoning development.