Skip to content
AI ConnectPowered by VELENTIS
AI-generated2 min

Race for the Riemann Hypothesis: AI-Assisted Proof Triggers Purely Human Counter-Draft

Levent Alpöge claims a Claude-driven Lean proof of the Riemann Hypothesis, prompting AMS President Ravi Vakil to rush out an entirely human alternative manuscript.

This article was AI-generated and published automatically. Context, labelling and all sources at the end of the article.

(KI-generiertes Symbolbild: Gemini / AI Connect)

An unprecedented competition has broken out in the mathematical community over one of the most prominent open problems in science. Number theorist Levent Alpöge, working at Anthropic, announced a formalized proof of both the Riemann Hypothesis and the Generalized Riemann Hypothesis. The proof was developed significantly with the assistance of Anthropic's Claude models and implemented in the Lean interactive theorem prover, making its logic directly verifiable by computer.

The reaction from the academic establishment arrived within hours. Ravi Vakil, a professor of mathematics at Stanford University and current president of the American Mathematical Society, rushed out a draft proof representing a human counter-effort. That work was authored by Stanford postdoctoral researcher Alessandro Masullo and had been reviewed over several months by mathematician Brian Conrad.

Vakil accompanied the announcement with an explicit note emphasizing that the work was entirely human, with no AI involved. The move highlights the acute tension within traditional mathematical institutions regarding machine-assisted discovery. For years, pure mathematics was regarded as relatively safe from generative models, which often struggled with precise logical reasoning and hallucinated formal steps.

Formulated in 1859 by Bernhard Riemann, the hypothesis governs the distribution of the non-trivial zeros of the Riemann zeta function and carries profound implications for the distribution of prime numbers. Listed among the Clay Mathematics Institute's Millennium Prize Problems, its solution remains a pinnacle goal of number theory. If Alpöge's formalization passes the Lean compiler without error, it would represent a historic milestone for machine-assisted mathematical deduction.

The sudden clash between Alpöge's formal AI approach and Masullo's non-AI manuscript underscores a broader cultural divergence in modern science. While machine learning advocates argue that large models paired with interactive provers scale human intuition, traditional mathematicians worry about losing deep intuitive comprehension of proofs. Both drafts will now undergo rigorous peer review and scrutiny across the global mathematical community.

What this means for you

This clash shows that AI-assisted formal reasoning has entered the territory of historic mathematical challenges. The rapid response from the American Mathematical Society reflects how deeply the frontier of AI is challenging conventional notions of academic credit and human intellectual primacy.

Evidence

Solidly sourced
54/100
  • Levent Alpöge announced a Lean-formalized proof of the Riemann Hypothesis and Generalized Riemann Hypothesis developed with Anthropic's Claude.

    single source
  • American Mathematical Society President Ravi Vakil reacted hours later by rushing out a rival draft proof by Stanford postdoc Alessandro Masullo.

    single source
  • The rival draft by Alessandro Masullo was proofread for months by Brian Conrad and explicitly labeled as all human with no AI.

    single source

The evidence score is computed, not hand-set: from confidence, the number of sources and the share of verified statements.

Source & transparency

As of: November 20, 2026

AI-generatedAI-generated: produced automatically from vetted sources with technical quality checks (source, quote and figure verification); no human sign-off of each item before publication

Sources
2
Verified statements
0 / 3
Evidence score
54Solidly sourced

Want to put this into practice?

We connect you with suitable AI providers from the DACH region, free of charge and without obligation.

What's next?