---
format: "aidr-story-markdown/v1"
id: "2e106d35b2bc0a07509e71b983e4d465accf78b1383050493a39d99ca5ce35ba"
canonical_url: "https://aidr.today/2e106d35?lang=en"
title: "25 Fields Medalists Warn AI Companies’ Benchmark Push Harms Mathematics"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-09-11T20:50:36.000Z"
category: "Research"
topics: ["openai","benchmark","reasoning","llm"]
source_urls: ["https://huggingnews.com/ai/update-25-fields-medalists-warn-ai-companies-benchmark-push-harms-mathem-2a0e4777","https://x.com/AndrewCurran_/status/2098493108224950438","https://x.com/7homaslin/status/2098479185022910676","https://x.com/AndrewCurran_/status/2098493117389549785","https://x.com/MTSlive/status/2098499849746583574"]
summary: "Twenty-five Fields Medal winners, including Terence Tao, published a joint declaration on Sept. 11 warning of a conflict between AI companies and the mathematics community. They wrote that “the push by AI companies to solve mathematical problems as a benchmark is detrimental to the science of mathematics, and to the mathematical community.” Tao has argued that efforts to solve mathematical problems advance the field by generating insights, not just answers. He warned that solving problems purely with AI, particularly without full transparency into the process, could become “a net negative for the progress of mathematics as a whole.” The declaration follows OpenAI’s Sept. 8 claim that about 10,000 AI agents using an unreleased model produced a proposed solution to the Navier-Stokes Millennium Prize Problem in 88 hours. OpenAI withdrew its sponsorship of the Caltech Mathathon on Sept. 10 after considering concerns raised by mathematicians, saying it wanted to engage more with the community on integrating AI and communicating its impact."
---

# 25 Fields Medalists Warn AI Companies’ Benchmark Push Harms Mathematics

> [Open the canonical story](<https://aidr.today/2e106d35?lang=en>)

**Published:** 2026-09-11T20:50:36.000Z
**Category:** Research
**Topics:** openai, benchmark, reasoning, llm

## Summary

Twenty\-five Fields Medal winners, including Terence Tao, published a joint declaration on Sept\. 11 warning of a conflict between AI companies and the mathematics community\. They wrote that “the push by AI companies to solve mathematical problems as a benchmark is detrimental to the science of mathematics, and to the mathematical community\.” Tao has argued that efforts to solve mathematical problems advance the field by generating insights, not just answers\. He warned that solving problems purely with AI, particularly without full transparency into the process, could become “a net negative for the progress of mathematics as a whole\.” The declaration follows OpenAI’s Sept\. 8 claim that about 10,000 AI agents using an unreleased model produced a proposed solution to the Navier\-Stokes Millennium Prize Problem in 88 hours\. OpenAI withdrew its sponsorship of the Caltech Mathathon on Sept\. 10 after considering concerns raised by mathematicians, saying it wanted to engage more with the community on integrating AI and communicating its impact\.

## Sources

- [Story source](<https://huggingnews.com/ai/update-25-fields-medalists-warn-ai-companies-benchmark-push-harms-mathem-2a0e4777>)
- [Story source](<https://x.com/AndrewCurran_/status/2098493108224950438>)
- [Supporting source](<https://x.com/7homaslin/status/2098479185022910676>)
- [Supporting source](<https://x.com/AndrewCurran_/status/2098493117389549785>)
- [Supporting source](<https://x.com/MTSlive/status/2098499849746583574>)