A gathering of roughly 100 leading mathematicians at the University of Cambridge spotlighted how rapidly advancing AI systems are beginning to assist with one of the field’s most exacting tasks: turning human-written arguments into machine-checkable proofs. This “formalisation” process, long seen as laborious, is benefiting from new AI tools that can translate and structure mathematical reasoning for proof assistants, raising hopes of faster verification and fewer errors. The shift could reconfigure mathematical workflows, enabling researchers to offload routine checking to computers and focus more on ideas and strategy. While enthusiasm is building, veterans caution that the technology must prove reliable at scale and integrate smoothly with established practices before it reshapes the discipline.





























