Artificial intelligence is widely used in clinical diagnosis, medical treatment, medical rehabilitation, disease prediction, and medical research in the medical field. However, due to errors in artificial intelligence technology, cases that damage the legitimate rights and interests of patients have occurred from time to time. Therefore, the application of medical artificial intelligence faces many risks. For example, the legal subject qualifications of medical artificial intelligence are not yet clear; controversies occur in the determination of legal liability when there are adverse consequences; there exist data bias and discrimination, and data fairness is difficult to guarantee; data privacy and personal information protection are insufficient; the legislative supervision system is not perfect; and medical artificial intelligence talents are not enough. In the face of these risks, it is necessary to define clearly the legal status of medical artificial intelligence, specify the responsibility of artificial intelligence medical accidents, select appropriate data samples and R&D staff to avoid prejudice, promulgate personal information protection laws as soon as possible, improve artificial intelligence-related legislation, strengthen comprehensive process supervision, and vigorously cultivate medical artificial intelligence talents, so as to promote the healthy and rapid development of medical artificial intelligence technology.