AI rewires math workflows

AI that can generate proofs and computations will not just speed traditional steps; it will change which steps matter, how mathematical work is modularized, and what roles human mathematicians play. That shift means productivity gains will be uneven — some tasks collapse, others (sense‑making, conceptual synthesis, writing coherent explanations) become the new bottlenecks. — If true, universities, journals, and funders must rethink training, authorship, peer review, and evaluation to accommodate a transformed research production model.