Generative AI can support mathematics teacher learning by automating parts of teacher rehearsals, including opportunities for teachers to practice specific mathematics teaching skills, receive feedback, and refine their approaches.
However, little is known about how to design generative AI solutions that incorporate the expertise of mathematics teacher educators. This article proposes a three-part design framework for integrating that expertise into generated feedback.
First, the framework builds on the modeling logic from evidence-centered design to develop design templates that capture the complexity of mathematics teaching. Second, it translates the expertise from these templates into prompts for generative AI.
Third, the article describes how a multi-agent system can use generative AI to simulate teacher rehearsals and feedback while preserving the complexity of educator expertise by splitting it across agents.
The authors explore how this framework could serve as a general approach for designing simulations that support complex, interactive rehearsals intended to enhance teacher learning.