With the wide application of new technologies such as large language models and generative artificial intelligence (AI) in the health care sector, artificial intelligence-assisted health care is confronted with new forms of legal risks. The algorithmic bias and data security issues in AI-assisted health care have given rise to risks of infringement on general personality rights and specific personality rights. The handling of health care data and the distribution of profits from health care data have spawned disputes over data property rights. Moreover, there will also be risks of uncertainties in the attribution of liability for medical harms once AI technology becomes deeply embedded in health care. Based on the emerging changes in the legal risks associated with AI-assisted health care, it is necessary to establish a corresponding algorithm review mechanism to eliminate algorithm biases, improve the data management system through a whole-life cycle approach to ensure data security, define hierarchical data property rights and establish authorization rules to resolve property rights disputes, and reasonably assign tort liability for medical harms based on specific faults.