Abstract: Traditional medical education assessment still faces issues such as insufficient diversification, weak formative assessment and feedback mechanisms, and the need to improve individualization, rigor, and precision. The in-depth integration of Artificial Intelligence (AI) technology is gradually revolutionizing the traditional assessment model, providing an innovative approach to enhance the objectivity, real-time performance, and personalization of assessment. This paper reviews the current application of AI in medical education assessment, including its advantages in empowering the diversification of assessment subjects, contextualizing skill assessment, personalizing process evaluation, refining assessment data, and intelligentizing assessment management. Meanwhile, it sorts out the practical cases of applying AI in scenarios such as multimodal assessment, laparoscopic operation evaluation, surgical video recognition, virtual standardized patients, and unmanned examination systems at The First Affiliated Hospital of Sun Yat-sen University. The study proposes an intelligent ″human-machine collaboration″ innovative ecosystem for medical education assessment centered on ″Multimodal Assessment-Accurate Feedback-Interactive Reflection-Need-Based Guidance″, which provides medical students with efficient and comprehensive dynamic assessment and real-time feedback, thereby facilitating the cultivation of their systematic thinking, critical thinking, and digital intelligence competence.

Key words: Artificial intelligence, Medical education assessment, Digital transformation, Human-machine collaboration, Innovative ecosystem