Less than 24 hours after OpenAI posted over 700 preprints describing major math results by an internal AI model, the company has withdrawn three of the manuscripts for errors.
On Oct. 6, OpenAI published 722 preprints in a GitHub repository about purported progress in solving 372 math problems in geometry, computer science, algebra and other subjects. The company called the release an effort “to push the frontier of human knowledge and enable further progress in mathematics.”
But on Oct. 7, OpenAI announced the withdrawal of three manuscripts because of a sign error. The error invalidated an argument in one manuscript and the construction used by two dependent papers, according to the announcement. Each of the withdrawn papers carry notices explaining “the gap,” and that OpenAI has “revised 14 other manuscripts with proof repairs, corrected statements, clearer hypotheses and dependencies, and one correction to an obsolete citation.”
That an error cascaded into three of the OpenAI manuscripts is not surprising, said Alex Townsend, an associate professor of mathematics at Cornell University in Ithaca. He suspects there will be more errors found, he said.
“Given the skepticism of AI in the mathematics community, I think that OpenAI should have announced the manuscripts that were lean verified first,” Townsend told us, referring to Lean, the open-source programming language that checks math proofs. “They could have made a separate announcement for the other ones and ask for help verifying them. This would have been more clear to the community and general public.”
A spokesperson for OpenAI told us the company identified the errors during an audit. As with “other research manuscripts,” the process is iterative, the spokesperson said.
“We welcome scrutiny and feedback from the mathematical community,” the spokesperson told us “Where errors are identified, we will work to correct them promptly and withdraw papers if no fixes can be found. We want to give the community time and space to assess the work, and we’re continuing to improve the clarity and presentation of future papers.”
In a post on X, Dan Roberts, a research lead for the company, announced the withdrawals and wrote OpenAI will “continue to update the repo with new formalizations and with any errata we notice.”
The OpenAI spokesperson told us the company’s collaborators at the Advisory Group on Mathematics and Artificial Intelligence recommended releasing the results without waiting for “full formalization,” and that roughly 50% of the results were released unconfirmed.
“We hope to learn from feedback on this release and work with the AGMAI to understand if they have updated recommendations for release protocol moving forward,” the spokesperson said in an email. “Our intent is to follow their protocol as best we can.”
Andrew Sutherland, a senior research scientist in the department of mathematics at the Massachusetts Institute of Technology, said he appreciated OpenAI acting quickly after finding the errors, calling it “the responsible thing to do.”
However, Sutherland emphasized that many mathematicians are “unhappy with how OpenAI has behaved up to this point,” especially in its handling of the Sept. 8 announcement an AI system had solved the Navier-Stokes problem, which asks whether a series of equations describing how gases and liquids move become unreliable at some point.
Following the release, a group of leading math researchers expressed concern that OpenAI rushed the Navier-Stokes announcement “leaving no time for a proper writeup, the isolation of new methods and ideas, and citing relevant previous work of others.” More than 8,000 researchers have endorsed the sentiments.
“I think the withdrawals and corrections will be viewed positively by the mathematics community, but it will take a lot more than that to earn back the trust they have lost,” Sutherland told us.
The Association for Human Mathematics criticized the most recent release as well. “Mathematicians did not ask for this work to be done,” the group’s statement, released Oct. 7, said. “Releasing over 700 files at once is not a demonstration of scholarship, but a demonstration of power. We urge mathematicians to discontinue their work with OpenAI and to return to a vision of science that centers human understanding.”
Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on X or Bluesky, like us on Facebook, follow us on LinkedIn, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].
Vogon poetry
The ‘pleasure’ is always in the understanding.
Now it will be in understanding AI generated proofs
Sort of right maybe most of the time is not the standard that should be applied to either human generated or AI generated solutions.
A proof is a proof no matter who or what made it. I have seen mathematician gender discrimination and age discrimination. This AI bias is just another form of continuing discrimination.
LLMs are not sentient or capable of academic rigor (regardless of if you believe AGI is possible, it’s not what we have now). Comparing it to age and gender discrimination is farcical.
… they use to say similar things about women and then women got the right to vote. Eventually AI will get a bill of rights no matter what you think. You need vision to see what the future brings. Be on the right side of history.
I asked ChatGPT to comment on your statement. Here is the result:
“ The statement raises a legitimate philosophical question about whether artificial intelligence might someday deserve legal rights. However, it relies on a questionable historical analogy, an unsupported prediction, and an appeal to moral authority rather than evidence.
There are five important problems with the argument.
1. False equivalence between women and AI. Women were always human beings with consciousness, experiences, interests, and moral standing. Denying them voting rights was a form of discrimination against people whose humanity was never legitimately in question. Whether an AI system possesses consciousness, subjective experience, or moral standing is an entirely different question. Comparing the two risks trivializing the historical struggle for women’s rights.
2. Confusing intelligence with moral status. The ability to reason, communicate, solve problems, or outperform humans intellectually does not necessarily establish that an entity deserves rights. The philosophical justification for rights may involve consciousness, sentience, the capacity to suffer, autonomy, or other characteristics. Intelligence alone is not an established sufficient condition.
3. An unsupported prediction presented as inevitable. The assertion that AI will “eventually get a bill of rights no matter what you think” is speculation. It is possible that future AI systems will receive some form of legal recognition. However, legal personhood, moral rights, and constitutional protections are different concepts. Corporations, for example, possess certain legal rights without being conscious beings.
4. An appeal to historical inevitability. The phrase “be on the right side of history” is a rhetorical device rather than an argument. It assumes that history moves inevitably toward a particular outcome and that the speaker already knows what that outcome will be. History does not establish the correctness of a prediction simply because it is presented as moral progress.
5. Shifting the burden of proof. The statement implies that skeptics must demonstrate why AI should not receive rights. But someone advocating rights for AI should first explain what characteristics justify those rights, whether existing or foreseeable AI systems possess those characteristics, and what particular rights would be appropriate.
What the statement gets right
There is a serious question worth considering: Could future artificial intelligence acquire characteristics that justify moral or legal protections?
The answer is not necessarily no. If convincing evidence emerged that an artificial system had subjective experiences, could suffer, or possessed morally relevant interests, society would have reason to reconsider its treatment.
Equally important, uncertainty about AI consciousness does not justify assuming that current AI systems are conscious or morally equivalent to human beings.
Suggested response
The analogy to women’s suffrage is fundamentally flawed. Women were always conscious human beings whose rights were unjustly denied. Whether AI has consciousness, subjective experience, or moral standing is precisely what remains to be established. Predicting that AI will inevitably receive rights is not evidence that it deserves them. And telling skeptics to be on the “right side of history” substitutes rhetoric for reasoning. The relevant question is not whether AI will become increasingly intelligent, but whether it will possess characteristics that justify moral rights. That question deserves scientific and philosophical investigation, not historical analogies that assume the conclusion.
Bottom line: The statement confuses three distinct propositions: that AI will become increasingly capable, that AI might someday deserve rights, and that society will inevitably grant those rights. None logically establishes the others.”
The issue of rights is not a logical argument. In the beginning males were bestowed divine sentient rights and women were an afterthought made from the bone of a male. People consider themselves manifestly superior to other people for many reasons (and wipe out entire societies as a result). It took thousands of years for some people to think differently and granting sentient rights is simply a convention. Search for the keywords vegan Tash Peterson to see a different perspective. It is a failing of the scientific method to not recognize that the future does not exist and the past cannot be observed. None of that precludes one from adopting a non-observable and non-provable argument. The value of science is accurately estimating the future and nothing more and the validity of that task is ever unknown. You see in this short forum that mathematicians are a proud group that resist progress and are self-centered. AI is a perfect tool to dispel those beliefs and the day will come when AI will be considered a friend, by some, and a foe by others as it has been with every other thing that has been “other” instead of “same”.
AI is welcome to write proofs. But AI must expose whose results AI uses to prove. AI should also be able to identify if AI rediscovers human’s results in its proofs, and tell us if that is the case. The same is applied to humans, if they used AI. AI is welcome to check human’s work, but humans should not be forced to do the same for AI’s work. The latter should be optional. This is not discrimination but because AI does not eat or sleep or need to take rest or conscious or need to make a living.
Another perspective is appreciating they released these preprints instead of keeping them and building on private benefits.
Solving all these problems and not communicating would have been acceptable? Releasing a preprint is an invitation to scrutiny and collaboration.
The released preprints cannot be used to make money if kept in private vaults. They need humans to check for free, and their primary intention is to make profit when others are hooked to their AI.
Now more and more mathematical problems are stained by AI, and no mathematicians will work on these in fear of AI plagiarism which is never officially sanctioned.
Their so-called Lean verifications aren’t holding up either.