OpenAI's feud with mathematicians is only escalating

A significant tension is building between leading AI development companies, particularly OpenAI, and the global mathematical community. This simmering disagreement recently escalated with an open letter signed by twenty-five Fields Medal winners—an honor often likened to the Nobel Prize in mathematics—criticizing AI labs for allegedly threatening their intellectual work and the fundamental values of mathematical research. This follows earlier efforts, such as the Leiden Declaration in June, which also addressed AI's impact on proof verification and academic credit.

The Core of the Conflict: "A Severe Misalignment of AI in Mathematics"

The open letter, titled "A Severe Misalignment of AI in Mathematics," was published on September 11, 2026, on the blog of renowned mathematician Terence Tao, himself a Fields Medal recipient. Among the signatories are other highly respected figures in mathematics, including Maryna Viazovska, Pierre Deligne, and Yu Deng, spanning nearly five decades of Fields Medal history.

The letter argues that the rapid race among AI companies to solve famous mathematical problems is "detrimental to the science of mathematics, and to the mathematical community." The mathematicians contend that the objectives of AI companies and the mathematical community are currently "severely misaligned." They see this as part of broader "alignment issues impacting other scientific and creative professions, as well as the whole of society."

Key Concerns Raised by Mathematicians:

• Threat to Intellectual Work and Understanding: The primary concern isn't merely that AI can solve problems, but that it bypasses the human process of understanding, insight, and the formulation of new ideas. Solving a problem is often just a tool or proxy for achieving conceptual understanding. The "mass production at faster and faster pace of 'true/false' statements could destroy fertile ground instead of breathing life into new ideas," the letter states.
• Attribution and Plagiarism: AI-generated solutions are often announced quickly, leaving little time for proper write-ups, isolating new methods, or crediting prior human work. This raises "severe attribution and plagiarism questions." Mathematicians argue that AI models produce results without citing the human work they build on, leading to concerns about recognition and intellectual property.
• "Slop Mathematics": Critics accuse AI companies of engaging in "slop mathematics," where they announce that their models have solved difficult theoretical problems, leaving human mathematicians to verify, disseminate, or even discredit these claims without compensation or credit. This labor, they argue, "goes uncompensated, uncredited, and unacknowledged."
• Commercialization Over Research Integrity: The mathematicians believe that AI companies are driven by the commercial incentive to "overstate the capabilities of their products" and use mathematical research as an "advertising opportunity." This prioritizes competitive wins against rivals over the advancement of mathematics as a discipline.
• Dependence and Inequality: There is a risk that mathematicians might become dependent on proprietary AI technology and expensive computational resources to produce competitive results, leading to inequality among researchers.

Recent Flashpoints: Navier-Stokes and the Caltech Mathathon

The feud intensified significantly with two recent events. Firstly, OpenAI announced on September 8, 2026, that an unreleased AI model had solved the Navier-Stokes existence and smoothness problem, one of the seven Millennium Prize Problems with a $1 million reward. OpenAI claimed its system worked on the challenge for 88 hours, using thousands of AI agents in parallel, with an estimated cost of $15 million in compute. This announcement, however, quickly drew criticism.

Notably, NYU mathematician Tristan Buckmaster and Anthropic's Levent Alpöge had been working on related aspects of the Navier-Stokes problem and had used OpenAI's own Codex software in their research. Buckmaster publicly questioned how OpenAI arrived at its solution so quickly, alleging that an OpenAI researcher, Sebastien Bubeck, twice proposed removing Alpöge from authorship on a paper about OpenAI's Navier-Stokes result. While OpenAI denied that its researchers or AI agents accessed Buckmaster's specific work before public release, it admitted it "cannot rule out that de-identified data derived from their usage of our products helped improve our models."

Secondly, the controversy spilled into the academic world regarding a "Mathathon" scheduled for October 30 at Caltech. This event, backed by approximately $2 million in AI credits from OpenAI and Anthropic, aimed to have participants use LLMs to tackle open research problems. However, a group of current and former Caltech mathematicians published an open letter arguing that the event "is likely to have destructive impacts for the mathematical community" and warned that sponsoring labs "will take credit for the effort of talented undergrads." In response to the backlash, OpenAI's research lead, Dan Roberts, announced on September 10, 2026, that the company would withdraw its sponsorship, stating, "We recognize that the rapid progress of AI in mathematics is disruptive. We're looking to engage with the math community more on the best way to integrate this technology and communicate its impacts."

Broader Implications for AI and Scientific Research

This escalating feud goes beyond a single company or a specific mathematical problem. It highlights fundamental questions about the future of intellectual work across all scientific and creative fields in the age of advanced AI. The concerns raised by mathematicians echo those voiced by artists, writers, and other creators regarding intellectual property, fair compensation, and the role of human creativity when AI systems are trained on vast datasets of human-generated content.

The debate forces a re-evaluation of what constitutes "discovery" and "authorship" in a world where AI can generate plausible but potentially flawed proofs or solutions. The human element of understanding, intuition, and the collaborative process of mathematical research is seen as vital, and many fear that an overemphasis on AI's problem-solving capabilities risks eroding these core values. As Terence Tao noted, if AI companies prioritize speed and headline-making breakthroughs, researchers might become hesitant to share promising work, thereby harming the principles of open and reproducible science.

The "Leiden Declaration on Artificial Intelligence and Mathematics," released in June 2026, further emphasizes these points. Endorsed by the International Mathematical Union, it calls for mathematicians to confront how AI companies are using published research without consent, bypassing peer review, and threatening the integrity of proof and attribution. The declaration identifies several threats, including AI systems producing unreliable arguments, lack of proper attribution, and the risk of research priorities being shaped by AI's amenability rather than deeper significance.

The Path Forward: Collaboration or Conflict?

The mathematical community is not against the use of AI as a tool. Many acknowledge its potential to assist in hypothesis generation, pattern recognition, and even theorem proving. However, they advocate for responsible integration of AI, ensuring that human understanding remains the primary goal. Recommendations from the Leiden Declaration include individual mathematicians disclosing all AI tool use, retaining personal responsibility for results, refusing to grant authorship to AI systems, and carefully considering which tools to use based on alignment with ethical values.

The ongoing discussions underscore the need for clear guidelines, ethical frameworks, and potentially new models of attribution and collaboration between AI developers and academic communities. The challenge is to harness AI's immense computational power and problem-solving abilities without undermining the human ingenuity, rigorous verification, and collaborative spirit that underpin scientific progress. OpenAI's withdrawal from the Caltech Mathathon, while a direct response to community concerns, also signals a recognition that engagement with the mathematical community is crucial for the responsible integration of this disruptive technology.

Ultimately, the debate calls for society to consider what it truly values in intellectual work. Is it just the answer, or is it the journey of discovery, the shared understanding, and the human insight that drives progress? The mathematical community's stand serves as a crucial reminder that as AI capabilities grow, so too must our commitment to ethical considerations and the preservation of human intellectual endeavors.

Frequently Asked Questions

What is the main concern of mathematicians regarding AI, particularly OpenAI?

Mathematicians are primarily concerned that AI labs, like OpenAI, are prioritizing the rapid "solution" of complex mathematical problems for competitive advantage and headlines, often without proper attribution to prior human work, rigorous verification, or a focus on fostering human understanding and insight. They fear this trend threatens the integrity of mathematical research and intellectual property.

What is the "Leiden Declaration on Artificial Intelligence and Mathematics"?

The Leiden Declaration, released in June 2026 and endorsed by the International Mathematical Union, is an open letter from mathematicians urging the community to address how AI companies use published research without consent, bypass peer review, and threaten proof integrity and attribution. It outlines five key threats to the discipline posed by AI.

Why did OpenAI withdraw its sponsorship from the Caltech Mathathon?

OpenAI withdrew its sponsorship from the Caltech Mathathon after an open letter from current and former Caltech mathematicians criticized the event. They argued that the Mathathon, which encouraged using AI to solve problems, could have "destructive impacts" on the mathematical community and serve as unfair advertising for AI companies. OpenAI stated it recognized the "disruptive" nature of AI in mathematics and sought to engage more with the community.

What is "slop mathematics" as described by the mathematicians?

"Slop mathematics" refers to the practice where AI companies announce purported solutions to difficult theoretical problems, often without full verification or proper write-ups. This then compels human mathematicians to verify, disseminate, or discredit these claims, often without compensation or credit for their labor.