AI-Mediated Formative Feedback in Open-Ended Mathematics Tasks

This study examines how AI-mediated formative feedback supports conceptual reasoning and metacognitive processes in open-ended university mathematics tasks. Conducted within the PRIN project “AI&F – Artificial Intelligence & Feedback for Effective Learning,” the research analyzes how AI-generated prompts guide students from intuitive reasoning toward mathematically grounded explanations. Two open-ended tasks were implemented with pre-service primary and secondary teachers, who worked collaboratively, submitted written solutions, received AI formative feedback, and revised their responses iteratively. Qualitative analysis of interaction excerpts and written productions shows that AI feedback functioned as a dialogic trigger rather than a purely evaluative response, fostering conceptual clarification, generalization, and reflective discussion. Results indicate that pedagogically designed AI feedback, embedded in collaborative settings, can enhance formative assessment practices and support effective human–AI partnership in mathematics education.