Terry Tao Backs Mathematicians Boycotting OpenAI Over AI Math Proofs

Field Medalist and UCLA mathematics professor Terence Tao recently amplified growing tensions between the academic research community and artificial intelligence developers by reposting a statement from the Association for Human Mathematics calling for a boycott of OpenAI. The dispute centers on a massive release on October 6, when OpenAI published 722 AI-generated mathematical manuscripts on GitHub, encompassing 372 result groups across number theory, geometry, algebra, and theoretical computer science. The release, generated by an undisclosed internal model, claims to have tackled several longstanding open problems, including variants of the Riemann hypothesis, the four-dimensional Gouba conjecture, and the Unique Games Conjecture. OpenAI stated the model evaluated approximately 4,000 problems, with most results following a standardized generation pipeline requiring compute equivalent to roughly three hours of high-end inference per output. While OpenAI released Lean formalizations to verify logical consistency, only 42 percent of the primary claims have been formally validated to date. The company acknowledged that unverified proofs may contain errors, having already withdrawn three manuscripts and revised fourteen others following initial reviews. The sheer velocity and volume of the release have triggered widespread concern among mathematicians. Critics argue that the output strategy bypasses traditional peer review and fundamentally disrupts the academic ecosystem, where understanding a proof underlying intuition typically requires years of community engagement. Theoretical computer scientist Scott Aaronson and his wife, Dana Moshkovitz, highlighted the dense, often incomprehensible nature of the texts, noting that human scholars risk being reduced to proofreaders tasked with deciphering machine-generated drafts. This friction follows a series of prior controversies, including disputed claims on the Navier-Stokes equations in September and debates over research priority, which prompted the formation of the Advisory Group on Mathematics and AI to establish transparent publication standards. Tao decision to amplify the Association for Human Mathematics statement underscores a broader dilemma within the research community. While Tao has consistently advocated for integrating AI tools into mathematical exploration, he opposes the transformation of fundamental research into a corporate benchmarking exercise. His stance reflects a split within academia: some researchers urge engagement to shape AI development guidelines, arguing that isolation forfeits academic influence, while others warn that unchecked commercial release models could subordinate mathematical inquiry to proprietary algorithmic roadmaps. OpenAI has pledged funding to host academic workshops for result validation, yet many scholars question who will bear the long-term costs of digesting AI-generated knowledge. The incident marks a pivotal moment in the intersection of artificial intelligence and pure mathematics. As large-scale models rapidly outpace human comprehension, the field must urgently reassess its verification protocols, authorship norms, and research priorities. The coming months will likely determine whether AI functions as a collaborative instrument in mathematical discovery or becomes a dominant force dictating the pace and direction of theoretical progress.