A mathematician from the artificial intelligence company Anthropic claims to have discovered a proof for the Riemann Hypothesis, one of the greatest challenges in mathematics, in collaboration with its chatbot Claude. This claim has not yet been independently verified.
On the 20th (local time), Anthropic mathematician Levent Alforge stated in a four-page post on his website, "I believe I have discovered a proof for the Riemann Hypothesis in collaboration with Claude." He also claimed that this method resolves a broader range of the 'generalized Riemann Hypothesis'.
Alforge initially released only the proof written in the programming language 'Lean', which is used to verify mathematical arguments with computers. He explained that the full narrative, which exceeds 250,000 pages, is "massive and unreadable, and posting a huge, unreadable 'slop' (low-quality output generated by AI) is clearly inappropriate." He also mentioned that he does not fully grasp the related mathematics and plans to organize the proof properly for writing.
According to an AFP report, Brian Conley, an authority in the field of the Riemann Hypothesis, described it as "a monumental moment in human history" and "the biggest event in mathematics." However, he noted that he has not seen the manuscript and has only reviewed the post on Alforge's website, for which he has requested a copy. Conley remarked, "It looks quite complex. I don't think a human could have discovered this," and predicted that a simpler method would emerge.
Nestor Guillen, a visiting professor in the mathematics department at New York University, positively assessed Alforge's candid admission that he does not fully understand the proof. He stated that the collaboration with AI as described by Alforge shows AI as more of a 'tool' than a 'machine god'.
Before the announcement, Alforge consulted with Dario Amodei, CEO of Anthropic, and acknowledged that the timing of the announcement could come off as cynical. This announcement came the day after Anthropic's record IPO.
In August, Anthropic revealed that an unpublished version of Claude improved upon longstanding results related to the Riemann Hypothesis, but emphasized that it did not prove the hypothesis. Two months prior, competitor OpenAI announced that its AI model had solved another Millennium Prize Problem related to fluid equations. Alforge faced a controversy over credit while conducting related research with NYU mathematician Tristan Buckmaster, and OpenAI rebutted claims of wrongdoing.
The Riemann Hypothesis, proposed by German mathematician Bernhard Riemann in 1859, concerns the distribution of prime numbers. The Clay Mathematics Institute designated this hypothesis as one of the seven 'Millennium Prize Problems' in 2000, offering a reward of 1 million dollars for a solution to each problem. According to the institute's regulations, to be recognized as an official solution, it must be published in a peer-reviewed journal and withstand verification for two years.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.