AI generates hundreds of mathematical results

OpenAI recently published hundreds of previously unpublished mathematical results generated by artificial intelligence. These results span multiple complex disciplines and have sparked concern and disorientation within the scientific community.

Mixed reactions from researchers

While some researchers have praised the quality of certain advances, many have criticized the lack of clarity, absence of systematic computer verification, and gaps in bibliographic attributions. Renowned mathematicians have expressed concerns about the negative effects of generative AI on their discipline and the future of academic careers.

Massive publication on October 6, 2026

On October 6, 2026, OpenAI published approximately 400 AI-generated results across more than 700 manuscripts covering various mathematical disciplines. The sheer volume of documents caused significant concern among experts, with some calling the event "pure madness." Reviewing the table of contents and abstracts alone took nearly an hour for some academics.

Existing criticism of OpenAI's AI systems

This controversy follows existing criticism of OpenAI's AI systems for their handling of the Navier-Stokes equations. OpenAI's practices are highly controversial among mathematicians, who are indignant at the idea that the company's AI systems could use their work. The "Association for Human Mathematics" has condemned OpenAI's publication, with approval from Terence Tao, a Fields Medal recipient and prominent mathematician.

Verification and quality concerns

Verification issues and concerns about low-quality content have been raised. While some demonstrations include formalizations in the Lean language, less than half of the manuscripts benefit from this rigor. OpenAI acknowledges that only about 42% of the main results have formalized proofs, leading experts to fear the need to review content of potentially mediocre or erroneous quality.

Criticism of the publication model

The "Association for Human Mathematics" argues that mathematicians have a unique vision of progress rooted in historical and disciplinary considerations. They reject OpenAI's claim that this publication advances the field and urge mathematicians and the public to approach the publication model with skepticism. Publishing over 700 files at once, they say, demonstrates power, not scientific rigor.

Potential value and challenges

Despite the challenges, the scientific community acknowledges that the publication contains works of remarkable quality. Some mathematicians, like Jared Duker Lichtman, have identified dozens of results with potential significance, including advances in the Riemann hypothesis and the Kakeya conjecture in four dimensions. These are major questions that researchers have struggled with for decades.

Reusing existing methods

Other scientists, such as Álvaro Lozano-Robledo, have noted that OpenAI's publications largely rely on existing methods and techniques reused with ingenuity. The sudden release of these results has disrupted academic research, with some researchers reporting that their work, including grant applications and thesis projects, has been rendered obsolete overnight.

Widespread academic disruption

The impact has been particularly severe in areas such as combinatorics, probability, and theoretical computer science, with group theory also heavily affected. Mathematician Bartosz Naskręcki described the situation as "real chaos" that could lead to a collapse of university culture and wipe out most departments.

OpenAI's response and advisory group

OpenAI has stated it is improving how it shares results with the mathematical community and consulted an independent group of experts. However, it ignored the group's recommendation to stop testing its models on advanced mathematical problems. The advisory group had emphasized that OpenAI's models remain inaccessible to the scientific community.

Ethical concerns raised

The "Association for Human Mathematics" criticized OpenAI for disregarding the advisory group's position, calling it a complete contempt for scientific research standards. They argue that OpenAI's actions risk undermining the reliability, ethics, and public interest of mathematical research.