Purpose/Significance To sort out the current status of medical artificial intelligence(AI)risk research in China, and to propose coping strategies for related risks.Method/Process Research hotspots are identified by using bibliometric methods. Based on the technology-organization-environment theory and textual analysis method,the paper summarizes and proposes a theoretical framework for medical AI risk identification from 4 levels:environment, organization, technology, and individual.Result/Conclusion A total of 24 specific risks are identified and recommendations are made in terms of institutional system construction, organization risk control, scope of individual obligations, and technology application,etc.