Mathematics teachers using generative AI to pose math problems related to students’ interests

Recent advances in generative AI (GenAI) allow mathematics teachers to personalize tasks to their students’ out-of-school interests in areas like sports or music. This can allow teachers to better integrate their students’ unique experiences into the mathematics classroom. It also capitalizes on affordances of generative AI to create fine-grained mathematical tasks about specific topics. We examine a group of seven 7th-grade mathematics teachers designing personalized tasks using ChatGPT while engaging in practices of problem-posing. The teachers then gave these tasks to their students, received student feedback, and iterated on their approach. We examine qualitative and quantitative data to explore what teachers learned from the process of personalizing problems with generative AI, how this approach fit into their workflow and curriculum, and what they perceived as the quality and usability of the problems ChatGPT helped them write. We found our qualitative data suggested teachers learned about using generative AI, about problem-posing, and about creating a classroom culture that incorporated students’ interests, showing growth on multiple dimensions of AI-TPACK. The personalization process was seen by teachers as reasonably fitting into their workflow, although issues with time efficiency, the diversity and variability of students’ interests, and the difficulty of learning to use GenAI, were barriers. Finally, we found the generated problems could match students’ interests at a fine-grained level and teachers saw their potential to increase engagement; however, teachers cited issues with the realism and authenticity of the problems.