A mathematician employed by Anthropic has put forth a groundbreaking claim: he utilized the company's artificial intelligence chatbot, Claude, to construct a proof for the Riemann hypothesis. This mathematical enigma, first proposed by German mathematician Bernhard Riemann in 1859, has remained unsolved for 167 years, challenging generations of mathematicians.
Levent Alpöge announced on November 20 that he spent approximately six weeks collaborating with Claude to produce a detailed mathematical argument. The claim, if independently verified, would represent a historic milestone for both mathematics and the rapidly evolving capabilities of artificial intelligence.
Understanding the Riemann Hypothesis
The Riemann hypothesis is a conjecture concerning the distribution of prime numbers. These fundamental building blocks of number theory (like 2, 3, 5, 7, 11) exhibit patterns that have long defied complete explanation. Riemann proposed that these patterns could be understood through a mathematical function known as the Riemann zeta function.
The hypothesis specifically states that every non-trivial zero of this function has a real part equal to one-half. Its resolution carries immense significance, impacting not only number theory and mathematical analysis but also areas of computer science that rely on prime number properties. It is also one of the seven Millennium Prize Problems identified by the Clay Mathematics Institute in 2000, with a $1 million prize awaiting a correct solution.
Claude's Role in the Claimed Proof
Alpöge's account suggests a novel form of collaboration between humans and AI. Rather than simply using Claude as a calculation tool, he engaged the chatbot as a partner over several weeks to develop and refine complex mathematical arguments. The process reportedly involved generating a detailed argument and formal proof files examinable using Lean, a programming language designed for verifying mathematical proofs.
The released materials include a four-page mathematical note and a vast collection of Lean files, reportedly comprising over 175,000 source files. This extensive formalization is intended for rigorous, computer-assisted verification.
The Verification Process Ahead
While the Lean code reportedly compiles, indicating the formal proof system has checked the encoded argument, this is not the final scientific verdict. Independent mathematical verification remains the biggest hurdle. Researchers must scrutinize the underlying logic, ensure the formalised statement accurately represents the Riemann hypothesis, and confirm that no errors or hidden assumptions undermine the result.
The Clay Mathematics Institute's rules for the Millennium Prize require a proposed solution to be published in an eligible peer-reviewed publication and to withstand at least two years of scrutiny. For now, the Riemann hypothesis remains an open mathematical problem, awaiting comprehensive validation.
Timing and Broader Implications
The announcement on November 20 came just one day after Anthropic's US stock market debut, adding a commercial dimension to the scientific claim. While Alpöge presented the work as a mathematical development, the timing has drawn attention amid intense competition in the AI industry.
This claim, alongside a separate, independent proof attempt by Stanford researcher Alessandro Masullo, highlights a growing interest in AI's potential for advanced mathematical research. Alpöge's assertion that Claude made particularly striking contributions and even outpaced him towards the end suggests a future where AI systems could become increasingly sophisticated research collaborators, exploring mathematical approaches that might take human researchers years to develop. However, human judgment, independent verification, and reproducibility remain essential.