A generative AI teaching assistant for personalized learning in medical education

Medical education faces a scalability crisis, where rising class sizes strain individualized instruction, while students increasingly adopt unvalidated Generative AI (GenAI) tools for individualized learning support. This study investigated how medical students integrate constrained GenAI systems into their self-directed learning practices using Retrieval-Augmented Generation (RAG), which limits large language model responses to instructor-curated materials, thereby reducing hallucinations while maintaining pedagogical utility. We deployed a RAG-based teaching assistant in a medical school basic science course across two consecutive cohorts, examining usage patterns, conversation content, and student feedback to understand adoption and learning behaviors. Students demonstrated strategic, context-dependent usage, with engagement intensifying during high-stakes assessment periods and substantial after-hours utilization. Users primarily sought clarification on foundational concepts and valued the system's continuous availability and source-grounded responses. However, knowledge-base constraints that ensured accuracy also limited broader inquiries, creating tension between reliability and comprehensiveness that shaped how students incorporated the tool into their study routines. These findings provide empirical evidence of how medical students navigate constrained AI tools for self-directed learning, informing institutional strategies for integrating these technologies into pedagogical frameworks.