New frontiers in simulation-based teacher training have been unveiled with the advancement of artificial intelligence (AI). Integrating AI into virtual student agents increases the accessibility and affordability of teacher training simulations, but little is known about how preservice teachers interact with AI-powered student agents. This study analyzed the discourse behavior of 15 preservice teachers who undertook simulation-based training with AI-powered student agents. Using a framework of ambitious science teaching, we conducted a pattern analysis of teacher and student talk moves, looking for evidence of academically productive discourse. Comparisons are made with patterns found in real classrooms with professionally trained science teachers. Results indicated that preservice teachers generated academically productive discourse with AI-powered students by using ambitious talk moves. The pattern analysis also revealed coachable moments where preservice teachers succumbed to cycles of unproductive discourse. This study highlights the utility of analyzing classroom discourse to understand human-AI communication in simulation-based teacher training.

Virtual reality (VR) offers immense freedom in the design of virtual instructional environments, but little guidance exists on how to capitalise on this freedom. This article reports on a study exploring how audience familiarity influences public speaking anxiety (PSA) and performance in a presentation speaking task in virtuo and in situ. Questionnaire instruments were used to gauge the PSA, motivation, focus, and self-confidence of 10 undergraduate students who each presented in four different audience conditions across VR and real life. Presentations were transcribed to identify features of performance, including utterance fluency, and speaking breadth and depth. Outcomes indicated that an audience of computer-generated agents resulted in less PSA than an audience of photorealistic people familiar to the speakers. Additionally, presenting to an audience of strangers in real life induced the most anxiety, but the performance features of articulation rate, disfluencies, and frequency of silent pauses were significantly improved in this condition. The main contribution of this study is to show that presentations directed at virtual audiences exhibit less fluent speech in non-native speakers than speeches to a real audience.

This work-in-progress paper reports on the establishment of preliminary reliability for a domain-agnostic evidence-centered assessment model to measure computational thinking (CT) in an online virtual world. Preliminary reliability was established between two researchers through manually coding 800 minutes of recorded learning sessions and over 350 minutes of consultation. Participants were three adolescents diagnosed with autism spectrum disorder. Findings indicate an acceptable level of reliability between the two coders, opening the way to more extensive application of the model in future studies.