Multilingual Gen-AI simulations for social-emotional learning in teacher education: a cross-cultural pilot study

This study focuses on the development and evaluation of a multilingual GenAI-mediated simulation for social-emotional learning (SEL), examining both its structural design and participant experience across three countries and languages. Adopting an exploratory design research approach, we analysed (a) the architecture of the simulation prompt and (b) the end-of-semester group interviews with 20 in-service and 66 pre-service teachers regarding their experience of the simulation. The analysis reveals a gated human-AI co-design process and a role-bound runtime architecture that sustains uncertainty through constrained interaction and delayed facilitation. Participants reported emotional activation, professionally authentic tension, and an extended reflective window, alongside perceived constraints requiring further development. By examining enactment across diverse linguistic contexts, the study highlights the potential of multilingual GenAI simulations for SEL, suggesting a scalable approach to expanding access to practice-based, culturally responsive teacher education.