Designing, Scaffolding, and Coding Complex Human-AI Interactions and Collaboration Processes: a Case Study of Clinical Reasoning in Medical Education
With the emergence of agentic artificial intelligence (AI), learners no longer interact solely with peers, teachers, materials, or systems; they now increasingly engage with AI technologies that possess a degree of agency. These interactions can be complementary, collaborative, or synergistic, yet effective engagement often requires thoughtful design and real-world complexity. This trend places greater demands on educators, researchers, and developers, making the design, scaffold, and analysis of complex human–AI interactions an emerging and vital area of study. In this seminar, Dr Yizhou Fan will present a case study on cultivating clinical reasoning skills in medical education to illustrate how he and his collaborators design a multi-agent human–AI interaction system for highly challenging, complex, and cognitively intensive learning tasks. In addition, he will introduce a human–AI collaborative coding approach developed to address the challenges of interpreting and applying massive amounts of behavioural and dialogical data while learners interact with AI.
Dr. Yizhou Fan is an Assistant Professor and Research Fellow at the Graduate School of Education, Peking University, and an Adjunct Research Fellow at Monash University. Over the past decade, his research has focused on learning analytics, self-regulated learning, and the application of artificial intelligence in education. He has published more than fifty papers in both Chinese and English and led several research projects funded by the National Natural Science Foundation of China, the Society for Learning Analytics Research (SoLAR), and the Alibaba Foundation. Dr Fan has received multiple national and international honours, including the QS Reimagine Education Award and the Emerging Scholar Award from SoLAR. He is recognised for his pioneering work on measuring and scaffolding learning using multimodal learning analytics, as well as for advocating awareness of learners’ metacognitive laziness when interacting with generative AI.