Generative <scp>AI</scp> , virtual reality simulation and communication entrustment in competency‐based medical education

Abstract Competency‐based medical education (CBME) requires learners to demonstrate communication as observable professional practice rather than theoretical knowledge. Immersive virtual reality (VR) simulation provides authentic opportunities to rehearse challenging clinical conversations, while generative artificial intelligence (AI) offers new possibilities for communication training. This correspondence argues that AI should function as a bounded reflective scaffold rather than as a substitute for live communicative performance. AI can support preparation and post‐simulation reflection by helping learners explore alternative responses, clarify reasoning and strengthen reflective practice. However, during simulated or assessed encounters, communication should remain entirely learner‐enacted to preserve the validity of entrustment decisions. Positioning AI outside the live interaction maintains professional agency, ensures assessment is based on observable human performance and protects the integrity of competency‐based progression. Establishing clear curricular boundaries allows educators to integrate AI into communication training while preserving the educational purposes of CBME and sustaining accountable professional practice.