#762: Exploring Collaborative Teacher-AI Interaction for Rehearsing Responsive Teaching in Elementary Mathematics Education

This proposal introduces TeacherCollab, an AI-supported simulated professional learning community designed to cultivate early-career teachers' (ECTs) responsive teaching in elementary mathematics. Rather than simulating students, TeacherCollab positions AI as collaborative teacher colleagues who model diverse pedagogical lenses while reasoning around authentic students' work. Grounded in approximations of practice and computer-supported collaborative learning (CSCL) theory, TeacherCollab creates a multi-agent dialogue environment where ECTs interact with AI-generated personas that surface students' relational dynamics, mathematical strategies, and areas of struggle. A pilot study with 16 ECTs showed that TeacherCollab fostered engagement, reflection, and awareness of multiple perspectives on student learning.