Fields medalists' AI misalignment letter draws a rebuttal aimed at mathematics itself

An open letter from the top of the field

Twenty-five Fields medalists have signed an open letter titled "A Severe Misalignment of AI in Mathematics," whose central claim, as relayed by Lior Pachter on his blog, is that the objectives of AI companies and those of the mathematical community point in sharply different directions. The letter faults labs for racing out results without a proper write-up, without isolating the new methods and ideas involved, and without citing relevant prior work — while mathematics is meant to be organized around conceptual understanding and insight.

According to Pachter, the letter appears to have been prompted by OpenAI's announcement of a solution to the Navier–Stokes existence and smoothness problem. He also spots an irony: the signatories admit they published quickly, without a more consultative process, because they lacked time — a rush he compares to the behavior they criticize.

A rebuttal from inside the discipline

Pachter, a computational biologist who spent eighteen years in UC Berkeley's mathematics department, agrees that AI companies are misaligned with mathematics' deeper aims. His counterargument is that the mathematics community itself has never been aligned with them either. If the profession truly treated students and ideas as its most precious resources, to be nurtured with great care, he writes, its record would look very different — particularly in the PDE and fluid mechanics territory where Navier–Stokes sits.

The historical record he cites

Juliusz Schauder, of Leray–Schauder degree fame, earned his doctorate in 1923 and was denied university positions through antisemitism among mathematicians, teaching high school while producing serious research. After the Nazis took power, many colleagues he asked for help declined, and Princeton refused him an invitation. He was murdered by the Germans; his wife Emilia, who hid with their daughter in sewers for a time, was later killed in a concentration camp.

Olga Ladyzhenskaya proved global existence and uniqueness for the two-dimensional Navier–Stokes equations in 1958, using the inequality that now bears her name. That year, aged 36, she was shortlisted for the Fields Medal but passed over for Klaus Roth and René Thom. Pachter cites 1958 committee records showing the prize was not decided purely on merit: Friedrich Hirzebruch was dropped as seemingly not needing encouragement, and Alexander Grothendieck, judged the most talented after him, was left to win later. Fifty-six years passed before Maryam Mirzakhani became the first woman Fields medalist; of the new letter's twenty-five primary signatories, Pachter notes, only one is a woman.

Karen Uhlenbeck, the first woman to win the Abel Prize in 2019, was told after her 1968 doctorate that departments did not hire women and that women were supposed to go home and have children. MIT, Stanford and Princeton were interested in her husband but not her, citing nepotism rules she later could not find any trace of, and she described her eventual post in Illinois as leaving her out of place mathematically and socially.

Cathleen Morawetz, first woman to direct the Courant Institute, past AMS president and first woman mathematician to receive the US National Medal of Science, once raised the scarcity of women in mathematics before an AMS governing body; Saunders Mac Lane's reply, as Pachter recounts it, was that mathematics is a very difficult subject.

Berkeley, and the present day

Pachter then turns to his own department. Julia Robinson, of Hilbert's tenth problem, was barred from teaching mathematics there for thirty-five years on nepotism grounds, appointed only after her 1976 election to the National Academy of Sciences, and required to document to the personnel office what she worked on every day. Decades later the department employed Yuval Peres, around whom mathematicians eventually described publicly at least seven cases or reports involving junior women, some of whom reportedly avoided conferences or lectures to escape his advances. Three years after Pachter arrived, the department lost an NSF grant specifically over poor student mentorship, and he writes that a now-emeritus colleague still denies sexism ever existed in the field. He adds accounts of a graduate student in tears after being told she was stupid, a professor saying of a ninth-year student that he was doing nothing because the student was stupid, and Yitang Zhang being told by his advisor that no Chinese student is good.

The question of the signatures

Pachter also questions why the primary signatories appended "Fields medalist" in parentheses rather than an affiliation or email address. Prizes, he argues, encode the judgments of a small group about which problems, fields and styles of mathematics matter — recalling Erdős's view that Szemerédi deserved a Fields Medal but that the deciders were not much interested in combinatorics.

Why it matters

The letter is a rare collective intervention by mathematics' most decorated researchers into how AI is reshaping proof: headline announcements without documentation, method isolation or attribution. Pachter's rebuttal, which reached Hacker News's front page, reframes the argument — if AI is to be judged by how well it nurtures understanding and the people who produce it, that is a standard mathematics has itself repeatedly failed, especially toward women and outsiders. For AI labs, the episode signals that credibility in mathematics will hinge on write-ups, citation and community norms rather than raw results; for mathematicians, it is an uncomfortable reminder that the norms being invoked against AI have a checkered record of being honored at home.