Mathematicians meet to define where AI belongs in the discipline
A Caltech report from an American Institute of Mathematics workshop captures a field thinking beyond both hype and rejection.
Mathematics is deciding what it wants from AI
A workshop reported by Caltech brought mathematicians together to discuss how artificial intelligence might change research practice, education, and the questions the community chooses to ask. The significance is less about a single product and more about a discipline taking ownership of the conversation.
From tool choice to research culture
AI can help search examples, translate between representations, suggest lemmas, or organise a large literature. None of those uses removes the need for mathematical taste: deciding which pattern is meaningful, which definition is useful, and which problem deserves years of attention. The workshop framing recognises that adoption is a cultural and institutional question as much as a technical one.
A question for departments
Universities can move beyond vague statements by defining concrete practices: disclose AI assistance, preserve reproducible notebooks, teach verification, and protect space for students to develop independent problem-solving habits. Those policies make room for useful tools without confusing assistance with authorship.