Mini-Symposium: Generative AI, Large-Language Models, and Machine Learning in Physics Education Research

The design and development of intentional activities using generative Artificial Intelligence (AI) is needed as the ecosystem of physics education continues to grapple with the role and impact of AI on the teaching and learning of physics. In this talk, we present our motivation, approach and results of a series of synthesis exercises provided to students in an introductory classical mechanics course that uses AI as a scaffolding tool. We highlight how using AI in this series of activities helps students build their own big picture of the course, and supports the refinement of their understanding of the material before each exam. We discuss the benefits and challenges of the development and implementation of each part of the synthesis exercises. The latter consists of an iterative categorization of physics equations, a concept map building activity, free-responses exercises and personalized student feedback. Through a mixed-method analysis of student responses, we underline students' perceptions and engagement with these exercises during the course. Our goal is to contribute to the ongoing conversation about practical pedagogical activities that effectively harness generative AI in the physics classroom.