Generative AI in Teacher Preparation: A Mixed-Methods Study of the Classroom Coach Simulation

• In Event: AERA Poster Session Saturday 1:45 pm
  In Poster Session: Generative AI in Teacher Education: Innovations in Preparation, Practice, and Pedagogy

Abstract

Preservice teachers often enter fieldwork feeling under-prepared to navigate the social, emotional, and ethical complexity of real classrooms. Immersive simulations that include technologies like large-language-models (LLMs) and virtual-reality (VR) settings offer a scalable way to rehearse high-stakes decisions while centering empathy and relational ethics. In this paper, we report on results of two mixed-methods pilot studies we conducted with preservice music teachers (n = 44), using the Classroom Coach simulation. The GPT simulation yielded high ratings for engagement and authenticity, with a statistically significant 27% increase in reported teaching self-efficacy. Preliminary findings revealed gaps in cultural nuance, sustained impact, and teacher co-design, pointing to the need for a more participatory, theory-driven approach.

Authors

• Priya Sharma, Pennsylvania State University

• Jacob Holster, Pennsylvania State University

• Aaron Knochel, Pennsylvania State University

• Jose Sandoval-Llanos, Pennsylvania State University

• Fouz Aljameel, Pennsylvania State University