Anthropic Mathematician Claims AI-Assisted Proof of the Riemann Hypothesis
Levent Alpoge, a mathematician at Anthropic, says he has proved the Riemann Hypothesis with the help of the company's Claude AI model. The proof has not yet been fully independently verified, but early reviewers rate the chance of error as very low.
Levent Alpoge, a mathematician at the artificial intelligence (AI) company Anthropic, has announced that he solved the Riemann Hypothesis, one of the best-known problems in mathematics, with the help of the company's Claude AI model. The proof has not yet been fully independently verified, so for now it remains a claim rather than a confirmed result. If it survives scrutiny, it would point to a sharp rise in AI capabilities in mathematics.
The hypothesis was posed in 1859 by the mathematician Bernhard Riemann and is 167 years old. It states that the distribution of prime numbers along the number line is not random but follows a complex order.
Checks are still under way
Alpoge published the proof, produced by Claude, in the programming language Lean, which is designed for the precise checking of mathematical arguments. Independent experts have already begun analysing the code. Andrew Sutherland, a mathematician at the Massachusetts Institute of Technology (MIT), said after compiling and checking the code that the probability of error is "approximately zero". Researchers are nonetheless carrying out final checks to make sure the proof accurately reflects the original hypothesis and does not exploit any loopholes in the Lean verification system.
Solving the hypothesis will have no immediate practical effect on everyday life: it will not change banking systems or make the internet faster. Scientists nonetheless regard the achievement as a major breakthrough. Hector Pasten, a mathematician at the Pontifical Catholic University of Chile, compared it to a "Rosetta Stone" linking different branches of mathematics. Alpoge himself expressed both sadness over the changing role of human mathematicians and optimism, calling the development "the dawn of something much bigger across science".
Another attempt, without AI
A few hours after Alpoge's announcement, Ravi Vakil, president of the American Mathematical Society, pointed to a different piece of work. Mathematician Alessandro Masullo had prepared a 940-page draft proof of the Riemann Hypothesis several months earlier, without AI assistance. That work is not yet ready for publication, and errors have been spotted in it. In the introduction, Masullo left a message for fellow mathematicians: "We are still here. We are very much alive. We are not going anywhere."

