This study examines how a customized version of ChatGPT can support mathematics teaching, focusing on the concept of continuity for real-valued functions. The work is grounded in the Knowledge for Teaching Mathematics with Technology (KTMT) framework and looks at how a personalized AI tool can help teachers design effective learning tasks, link different mathematical representations, and connect exploration with rigorous justification. In-service mathematics teachers worked with both the default and a customized version of ChatGPT to create instructional materials. The evaluation combined a qualitative analysis of the resources with a quantitative analysis using a Fuzzy Cognitive Map (FCM) based on the KTMT framework. The qualitative results show that the customized model enabled teachers to produce richer, more varied, and pedagogically coherent materials, encouraging the use of multiple representations and promoting deeper conceptual understanding. The FCM analysis confirmed these trends, showing clear gains across all three KTMT dimensions. Teachers reported that the customized ChatGPT helped them move smoothly between algebraic, graphical, and tabular forms, supported exploratory approaches to problem-solving, and encouraged students to justify their reasoning. Overall, the findings suggest that when aligned with a structured pedagogical framework, AI tools like ChatGPT can enhance mathematics instruction and support teacher development.
ChatGPT for Innovative Mathematics Teaching: A Mixed-Methods Study
Maria Lucia Bernardi;Roberto Capone;
2026-01-01
Abstract
This study examines how a customized version of ChatGPT can support mathematics teaching, focusing on the concept of continuity for real-valued functions. The work is grounded in the Knowledge for Teaching Mathematics with Technology (KTMT) framework and looks at how a personalized AI tool can help teachers design effective learning tasks, link different mathematical representations, and connect exploration with rigorous justification. In-service mathematics teachers worked with both the default and a customized version of ChatGPT to create instructional materials. The evaluation combined a qualitative analysis of the resources with a quantitative analysis using a Fuzzy Cognitive Map (FCM) based on the KTMT framework. The qualitative results show that the customized model enabled teachers to produce richer, more varied, and pedagogically coherent materials, encouraging the use of multiple representations and promoting deeper conceptual understanding. The FCM analysis confirmed these trends, showing clear gains across all three KTMT dimensions. Teachers reported that the customized ChatGPT helped them move smoothly between algebraic, graphical, and tabular forms, supported exploratory approaches to problem-solving, and encouraged students to justify their reasoning. Overall, the findings suggest that when aligned with a structured pedagogical framework, AI tools like ChatGPT can enhance mathematics instruction and support teacher development.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
