“Proof, Plausibility and Pedagogy: Rethinking Mathematical Thinking with Generative AI.”

Generative AI can produce fluent solutions and mimic mathematical argument, which forces mathematics education to confront a new boundary between doing mathematics and delegating it. This talk examines how AI shifts the centre of gravity from producing answers towards judging validity and making informed choices about methods and tools. It covers how classroom tasks can treat AI output as a stimulus for critique and refinement so that students practise verification and sense making rather than copying polished responses. The talk connects these task designs to assessment by highlighting approaches that capture reasoning and responsiveness to feedback even when AI is available. The session also includes practical workflows for teachers and lecturers, including prompt literacy for mathematics and ways to build student agency when mathematical work is increasingly co-produced with AI.