Terence Tao at ICM: AI Ushers Math into an Era of Proof Surplus

At the International Congress of Mathematicians (ICM 2026) in Philadelphia, mathematician Terence Tao delivered a public lecture titled "Mathematics in the Age of AI" and participated in an "AI for Mathematics" panel discussion alongside math professors and Google DeepMind researchers. He systematically articulated AI's profound impact on mathematics, noting that the field is shifting from an era of "proof scarcity" to "proof surplus," with the very foundations of the discipline facing severe turbulence.

Confirmed

Based on on-site slides and attendee accounts, Tao broke down mathematical work into a five-stage pipeline: proof generation, proof verification, presentation, publication, and canonization (integration into textbooks/classics). He assessed that AI is currently very good at the first step (generation) and is making progress on the second (verification), but the subsequent stages remain highly dependent on humans. In practical applications, AI can quickly synthesize massive numbers of papers and brute-force test ideas, while humans often only need to check a few examples to spot patterns. Additionally, he listed core goals of mathematical research, including solving open problems, developing new theories and methods, understanding the world, building the mathematical community, and training the next generation. Interestingly, Tao specifically noted that the em-dashes in his presentation slides were "human-generated."

Why it matters

Tao argued that the foundational "values and practices" of mathematics are facing a crisis. Once intelligence becomes cheap, merely "solving a problem" no longer automatically generates value; true value lies in whether the solution can be utilized by humans. Although this "human-machine efficiency gap" is currently obvious, it may gradually narrow in the future. He emphasized that if the mathematics community can thoroughly examine and adapt to these changes, the turbulence brought by AI could transform into new opportunities for disciplinary growth. Furthermore, this impact is not limited to mathematics; purely theoretical disciplines like economics are facing similar crises, such as job market difficulties and a decrease in pure theoretical research in top journals.

2026-07-25 ~ 2026-07-26 · 14 related posts

• [source] Terence Tao warns mathematics is entering a turbulent age in the era of AI — AlexKontorovich · 2026-07-25
• Terence Tao’s ICM 2026 slide lays out what math is for in the age of AI — AlexKontorovich · 2026-07-25
• Terence Tao’s ICM 2026 talk says AI could move math from proof scarcity to abundance — AlexKontorovich · 2026-07-25
• Terence Tao says AI can brute-force math ideas, but the gap may shrink — rohanpaul_ai · 2026-07-25
• [source] Terence Tao says AI is pushing math from proof scarcity to proof abundance — lmthang · 2026-07-25
• Terence Tao Shares ICM 2026 Slides Amidst AI-Driven Crisis in Academic Theory — Afinetheorem · 2026-07-26
• Terence Tao joins an AI for Mathematics panel at the International Congress of Mathematicians — Thom_Wolf · 2026-07-26
• [source] ICM2026 talk frames mathematics as a five-stage pipeline AI only partly covers — lmthang · 2026-07-26
• Terence Tao talk prompts a deeper question: what counts as solving a problem in the AI era — samiramanabi · 2026-07-26
• Terence Tao discusses mathematics and AI, with a note that the em dashes were human-generated — juliusadml · 2026-07-26
• Terence Tao discusses mathematics in the age of AI — burny_tech · 2026-07-26
• Terence Tao frames AI in mathematics as a testable capability conjecture — burny_tech · 2026-07-26

2 near-duplicate retellings: soumitrashukla9 · burny_tech