Abstract: This study systematically explores the application of generative AI technology in the basic biomedical course ″Anatomy and Physiology″, as well as the risk control mechanism. The goal is to develop a teaching system that is ethically sensitive, logically structured, and effective. First, the study developed a ethical risk identification and evaluation framework based on the GB/T 27921-2023 standard, which clarified the main ethical risk sources in the classroom, such as data leakage, information bias, ability degradation, and teacher weakening. This framework also innovatively created an ethical risk matrix, providing a quantitative tool for AI governance in medical education. Second, based on the three-dimensional integration concept of ″knowledge dimension-ability dimension-ethics dimension″, the study proposed a teaching strategy system that is structured, visible, capable of evaluation, and ethical. This system significantly improves students' three-dimensional recognition accuracy and enhances their critical awareness of technology. Finally, the study developed a virtual and physical fusion electric physiology experiment platform that combines intelligent bodies, knowledge graphs, and AI-assisted case libraries. This platform saves experimental resources and optimizes clinical thinking training. The results of this study demonstrate strong practical value, real-world relevance, and potential for widespread application, providing replicable exemplary support for the digital transformation and intelligent development of medical courses.

Key words: Artificial intelligence, Generative, Anatomy and physiology, Ethical risks, Three-dimensional integration, Virtual simulation teaching