Using AI chatbots to facilitate mathematics preservice teachers’ noticing skills

The ability to notice and interpret relevant aspects of classroom interactions is central to effective teaching and plays a critical role in teachers’ professional development. This study investigated the development and testing of an AI chatbot (NiCo) designed to support preservice teachers’ noticing skills. The chatbot was built using the GPT-4o model and included a meta-prompt that focused on structured support and providing feedback based. The chatbot was tested with 25 mathematics preservice teachers analyzing classroom videos in the field of secondary algebra education. Data collection involved questionnaires on prior experience with the content, self-assessed AI competencies, and user perception of the chatbot. Results indicate that preservice teachers experienced the chatbot as motivating and helpful for developing their noticing skills. They reported support in perceiving classroom events, interpreting them, and generating alternative actions. Usability was rated positively, though improvements were suggested regarding input options, accessibility, and subject-specific guidance. The study demonstrates the potential of AI chatbots to provide individualized support for the development of teacher noticing skills in teacher education. With the empirically validated meta-prompt structure of NiCo, we propose a novel, accessible and scalable format for AI-enhanced teacher education and even professional development. By demonstrating both feasibility and positive user perceptions, this research contributes concretely to the academic discourse on AI-driven interventions for preservice teachers’ noticing skill development.