BACKGROUND:
Medical knowledge systems are complex and rapidly evolving. Medical students face challenges including excessive knowledge volume, fragmented structures, and limited interdisciplinary integration.Traditional textbook and lecture based teaching models are insufficient for supporting deep understanding and continuous knowledge updating. Generative artificial intelligence offers new opportunities by integrating multi-source medical knowledge, constructing structured knowledge systems, and providing contextualized learning support. Meanwhile, with AI increasingly applied in clinical practice, medical students must develop not only AI-assisted learning skills but also competencies in understanding, evaluating, and responsibly applying AI. Current medical education often relies on isolated AI tools, lacking systematic platforms that integrate medical knowledge learning with AI literacy development and assessment.

METHODS:
Based on the Guangdong Medical University AI Medical School (GDMU-AIMS), A generative-AI-powered teaching–assessment integrated platform was developed. The platform adopts a parallel architecture of disease-specific large-model training and knowledge-base–driven applications. Expert-developed disease-specific large models enable free question–answer interaction and inquiry-based learning within targeted disease domains. In parallel, authoritative medical resources—including clinical guidelines, expert consensuses, textbooks, regulations, and national examination syllabi and item banks were integrated into a traceable, updatable structured knowledge base. Retrieval-augmented generation supports virtual patients, medical knowledge learning, and examination training, while embedding AI literacy cultivation and competency assessment.

RESULTS:
By integrating authoritative medical knowledge graphs, the platform established a clinical decision support system featuring guideline querying, free interaction, and privacy protection. It provides guideline based screening recommendations, interpretation of abnormal results, standardized referral guidance, and health education support. Privacy preserving technologies ensure secure handling of sensitive data and anonymized interactions. The platform offers an open, interactive, and immersive learning environment that supports autonomous learning and the development of clinical reasoning.

CONCLUSIONS:
GDMU-AIMS enhances learning efficiency and it supports flexible, interactive, learner-centered medical education and provides a scalable model for training AI-augmented medical professionals in the AI era.