Sunday 22 November 2026 Live: Iran war Stories Browse Wire Subscribe
Faultline Journal — home

TechnologyAnalysis

A Claimed Proof of the Riemann Hypothesis, Awaiting Scrutiny

An Anthropic mathematician says Claude helped him prove a 160-year-old conjecture, but no independent review has yet taken place.

A Claimed Proof of the Riemann Hypothesis, Awaiting Scrutiny

Reporting drawn from SCMP World. Analysis and editorial framing by the Faultline Journal Editorial Desk.

Until mathematicians have read and tested the argument, the appropriate description is a claim, not a result.

A mathematician at the AI company Anthropic said on Friday that he had used the firm's Claude chatbot to prove the Riemann Hypothesis, a problem that has resisted mathematicians for more than 160 years. According to the report, Levent Alpoge wrote in a four-page note released on Friday: "I believe I have, working with Claude, discovered a proof of the Riemann Hypothesis." The claim, as the report makes clear, has not been independently verified, and the proof has not yet been reviewed by outside experts.

The Riemann Hypothesis, first proposed by the German mathematician Bernhard Riemann, concerns the distribution of prime numbers and is widely regarded as one of the deepest open questions in mathematics. A valid proof would carry consequences across number theory and beyond. That is precisely why the claim demands more than a press cycle: in mathematics, a proof becomes a proof only when the community has checked it, and four pages is a short document for a result of this scale.

The involvement of a commercial AI system adds a second layer of uncertainty. The report does not say how Claude contributed, whether it generated steps, checked them or served as a sounding board, and it offers no detail on the structure of the argument. Nor is it known whether the note has been submitted to a journal or posted to a preprint server for wider examination. Until mathematicians have read and tested the argument, the appropriate description is a claim, not a result.

The episode nonetheless matters. It shows that leading AI firms are now willing to attach their names to claims at the frontier of pure mathematics, and that researchers inside those firms are using their own tools in earnest. If the proof survives review, the question of how credit is assigned between human and machine will become urgent. If it does not, the episode will be a caution about announcing breakthroughs before verification. Either way, the next move belongs not to Anthropic but to the mathematicians who must read the note.

Riemann HypothesisAnthropicClaudemathematicsAI research