#571: Preservice Teachers’ Dialogic Talk Moves and Beliefs About Control in a AI-Driven Virtual Reality Classroom Simulation

This paper presents two studies investigating preservice teachers’ dialogic teaching practices using an immersive, large language model (LLM)–enhanced virtual reality (VR) classroom simulation environment. Study 1 examined 23 preservice teachers’ perceptions of authenticity, situational interest, and self-efficacy during VR-based classroom debates and analyzed the talk move pattern using hidden Markov model. Participants perceived the simulation as authentic and engaging but relied predominantly on superficial probing questions rather than dialogic moves that deepen students’ thinking. Study 2 explored whether repeated simulation experiences combined with peer feedback influenced beliefs about classroom ownership of ideas and activities among 22 preservice teachers. Mixed-effects modeling indicated significant belief shifts toward student ownership of ideas, moderated by engagement with the simulation. Collectively, the findings suggest that AI-driven VR simulations provide authentic, scalable environments for developing dialogic competence and reshaping teacher beliefs toward more student-centered instructional practices.