Blog — Latest in Math Tech & AI
News, studies, deep-dives, and trends at the intersection of mathematics and AI.
An AI-discovered fast dynamo shows mathematical AI edging from benchmark work toward publishable analysis
Posted to arXiv on August 20, 2026, a new analysis-and-probability paper reports a smooth random fast dynamo on the three-torus and states that the central proof idea was generated autonomously by an AI system before human verification and writing.
Terence Tao asks what mathematics should protect as AI reaches research level
Posted to arXiv on August 17, 2026, Terence Tao’s new essay shifts the conversation from whether AI can do mathematics to what mathematicians should still value, teach, and reward when it can.
FAR shifts AI mathematics from solving one chosen problem to searching a whole research direction
Posted to arXiv on August 17, 2026, the FAR pipeline shows how AI can scan mathematical literature, surface open conjectures, attempt them at scale, and send only the strongest artifacts to expert review.
FaithformBench asks whether mathematical autoformalization tools stay faithful when reasoning goes wrong
Posted to arXiv on August 11, 2026, FaithformBench studies whether AI systems that translate mathematical reasoning into Lean preserve both valid and invalid steps instead of silently “fixing” the input.
FormaTheoria turns AI-assisted Lean formalization toward one of mathematics’ biggest long-range targets
Posted to arXiv on August 11, 2026, FormaTheoria presents an AI-assisted workflow for rebuilding large mathematical theories in Lean, using finite-group results tied to the Classification of Finite Simple Groups as a demanding test case.
MechGeo pushes AI geometry further by turning Olympiad diagrams into Lean-checked proofs
Posted to arXiv on August 3, 2026, MechGeo combines autoformalization, counterexample-guided repair, and kernel-checked proving to tackle Euclidean geometry problems in Lean 4.
BlueprintRepair turns failed Lean proof plans into smaller, cheaper AI repair jobs
Submitted to arXiv on July 30, 2026, a new formal-mathematics study shows that schema-checked local edits can repair failed Lean proof blueprints almost as effectively as free-form rewrites while using fewer tokens.
From lecture notes to Lean, a probability textbook becomes a checkable AI-era math resource
Submitted to arXiv on July 29, 2026, a new Lean formalization project argues that turning a probability textbook into machine-checked mathematics can strengthen both teaching and reliable AI-assisted proof work.
Nature warns that AI in mathematics needs rules before bad habits harden
Published on July 27, 2026, a new Nature World View argues that mathematics should set stronger norms for AI use now, before opaque tools, weak attribution, and low-quality machine-written papers become routine.
Showing page 7 of 11 · 92 total