Altman Responds to AI's Impact on the Mathematics Community: Understands Mathematicians' Situation, but OpenAI Will Not Stop Research

Altman Responds to AI Disruption: Understands Mathematicians' Situation, but OpenAI Will Not Stop Research

IT Home Download the Client for Exclusive Early Access October 30, 2026, 21:56:09 From Shandong Updated October 31, 2026, 07:19:20

IT Home, October 30 news: OpenAI CEO Sam Altman said on the "Cheeky Pint" podcast that he understands the situation mathematicians face as AI research results proliferate, but OpenAI needs to continue publishing scientific discoveries of significant value.

He believes that humanity has the right to understand the fundamental laws governing the universe, and that the company should not keep major basic scientific discoveries internal for long periods.

On the podcast, Altman acknowledged that OpenAI's previous approach to publishing mathematical research results had room for improvement. He said he hopes to advance scientific progress while respecting as much as possible the research communities that have driven the development of disciplines such as mathematics for centuries, and to help researchers in related fields adapt to the changes brought by AI.

This year, OpenAI has successively published a large number of mathematical research results produced by AI, sparking controversy in the mathematics community over how results are released, the quality of proofs, and academic attribution. Researchers have criticized that some of the company's papers lack sufficient background explanation, references, and clear argumentation, forcing mathematicians to spend large amounts of time verifying results, understanding proofs, and determining whether the relevant discoveries truly hold.

On the podcast, Altman revealed that OpenAI had intended to increase investment in mathematical research over a short period in order to explore the capability boundaries of AI models. He mentioned that the company has an internal Slack channel called "Model Name-Hold My Beer," where employees give the model the hardest problems from their respective academic careers to try to solve. He said the team did not think the model would be able to complete some of those tasks at the time, but the results exceeded expectations, even exciting him so much that he could not sleep.

However, Altman also said that when he saw AI could solve some research problems previously considered extremely difficult, he felt a stronger sense of loss and melancholy than expected. He pointed out that part of the value of mathematical research lies in the fact that researchers, through long-term effort, have the opportunity to solve the most challenging problems, and AI is changing that possibility.

By his estimate, the total resources OpenAI has invested in mathematical problem research amount to hundreds of millions of dollars, of which the work on solving problems related to the NS equations cost about $10 million (IT Home note: currently about 66.769 million yuan), and possibly more. He also said that as model capabilities improve and research methods advance, the cost of redoing similar work in the future may fall to about one percent of the current level.

OpenAI has also recently faced new controversy over how AI research results should be published. In October this year, the company announced three research advances in mathematics, computer science, and theoretical physics. Among them, some mathematics and computer science results have been made public through papers and formal proofs; the physics result involving Yang-Mills theory has not yet had complete materials released.

OpenAI said this release approach was adopted based on the recommendations of an external expert panel, but some researchers have questioned it, arguing that restricting public access to research materials may hinder independent verification and academic exchange.

In addition, some mathematicians previously opposed OpenAI's participation in mathematics-related activities at Caltech, arguing that AI companies using complex mathematical problems as indicators for testing model capabilities may have a negative impact on the mathematics research community. Twenty-eight Fields Medal winners also jointly published an open letter criticizing the practice of AI companies using the solving of difficult mathematical problems as benchmarks, saying it may harm the discipline of mathematics and its research community.

In response to these criticisms, Altman said OpenAI is exploring how to make scientific discoveries benefit the public while providing better development conditions for researchers in related fields. He stressed that he hopes to find a way of publishing results that is more respectful of the academic community, but believes humanity should not therefore give up progress in basic scientific research.

OpenAI previously also announced a $5 million donation (currently about 33.384 million yuan) to a new program of the Simons Foundation, to support mathematicians in studying, understanding, and further developing AI-generated results, including funding workshops, academic conferences, and related research projects.

"Special Statement: The above content (including videos, pictures, and audio if any) is uploaded and published by users of "Dafeng Hao," a self-media platform under Phoenix Net. This platform only provides information storage space services.

Notice: The content above (including the videos, pictures and audios if any) is uploaded and posted by the user of Dafeng Hao, which is a social media platform and merely provides information storage space services."