Terence Tao: AI's Impact on Math Research is Like Cars on Cities—Need New Infrastructure for Machines

Mathematician Terence Tao compares AI's impact on mathematical research to cars' impact on cities: cars are fast but conflict with old road systems, requiring new planning to integrate. Existing math systems (journals, mentorship, etc.) are designed for humans—human proofs are slow but build skills and discover new directions; AI-assisted proofs are efficient but lose these additional values, making them hard to meet traditional journal publication requirements.

Tao believes we should not force AI to fit old systems but create new math infrastructure for machines (like formal proof assistants, auto-generated rough proof libraries) and propose a new "AI planning" discipline. He notes AI enriches his work (e.g., charts, code, literature retrieval) but core research still uses pen and paper; AI reduces idea generation cost, but verification becomes the new bottleneck.