ChatGPT as an interactive learning environment: supporting creative problem posing in teacher education

Abstract

This qualitative multiple-case study examines how generative artificial intelligence can serve as an interactive learning environment to enhance the creative problem-posing skills of pre-service mathematics teachers. Twenty-four participants engaged in problem-posing tasks based on data tables, images, and open-ended scenarios. Their problems were evaluated across fluency, flexibility, and originality-the three core dimensions of creative thinking. Based on pre-test results, twelve participants with the lowest creativity scores entered a structured ChatGPT-supported intervention, guided by the orientation-connection-generation model. During this phase, participants revised their problems through dialogic engagement with ChatGPT, then generated new problems in a post-test. Comparative analyses revealed varied developmental trajectories: some participants reached fully developed levels across all dimensions, others sustained high performance with limited gains, and a smaller group showed selective improvement. Overall, ChatGPT contributed to increased productivity, greater representational diversity, and richer contextual originality. Participants also demonstrated growth in strategic thinking and pedagogical awareness, positioning AI as more than a suggestion generator. These findings highlight the potential of generative AI to foster creativity-oriented teacher education and offer practical implications for integrating AI-supported learning scenarios into instructional design.