Anthropic Researcher Claims AI-Assisted Proof of Riemann Hypothesis
Anthropic researcher Levent Alpöge claims to have proved the Riemann Hypothesis using the Claude AI model. The 17-million-line Lean code has passed initial verification by MIT’s Andrew Sutherland and others, who estimate error probability near zero. However, experts urge caution, noting that formal verification does not guarantee the proof addresses the correct problem. The claim, if valid, would solve a $1 million Millennium Prize Problem.
Key points
- Levent Alpöge, a researcher at Anthropic, claims to have proved the Riemann Hypothesis using the Claude AI model.
- The proof is written in Lean, a formal verification language, and consists of approximately 17 million lines of code.
- MIT senior research scientist Andrew Sutherland verified the code and estimates the chance of error is roughly zero.
- Other mathematicians, including Imperial College London’s Kevin Buzzard, urge caution, stating the proof’s implications are not yet understood.
- The Riemann Hypothesis is one of the Clay Mathematics Institute’s seven Millennium Prize Problems, each carrying a $1 million reward.
AI-Assisted Mathematical Breakthrough Claimed
Levent Alpöge, a researcher at the artificial intelligence company Anthropic, has announced that he has discovered a proof of the Riemann Hypothesis with the assistance of Claude, Anthropic’s large language model. The Riemann Hypothesis, first proposed by Bernhard Riemann in 1859, is a conjecture about the distribution of prime numbers and stands as one of the most famous unsolved problems in mathematics. Alpöge made the announcement in a note posted on his website on Friday, as reported by Bloomberg. He stated that his work also covers a broader version known as the generalised Riemann hypothesis.
Formal Verification via Lean Code
Rather than submitting a traditional academic paper, Alpöge released the proof in Lean, a programming language designed for formal mathematical verification. This approach requires every logical step of the proof to be checked by a computer, eliminating human error in deduction. The code generated for this proof is substantial, running to approximately 17 million lines, according to reports from Science. This scale underscores the complexity of the problem and the potential role of AI in managing vast logical structures.
Independent Verification and Expert Caution
The claim has already attracted attention from prominent mathematicians. Andrew Sutherland, a senior research scientist at the Massachusetts Institute of Technology (MIT), independently checked the code on his own computer. He reported that at least three other professional mathematicians had also examined the proof. Following these checks, Sutherland stated that the probability of any error in the proof is roughly zero, as reported by Scientific American. However, not all experts are ready to declare the problem solved. Kevin Buzzard, a professor at Imperial College London, warned that a proof compiling in Lean does not automatically mean it addresses the correct mathematical problem. He emphasized that humans must still confirm that the formal statement accurately reflects the Riemann Hypothesis. Buzzard also noted that nobody yet understands what the proof shows or its broader implications.
Significance of the Millennium Prize Problem
The Riemann Hypothesis is one of the seven Millennium Prize Problems established by the Clay Mathematics Institute. Each problem carries a prize of $1 million for a correct solution. A verified proof would not only resolve a central question in number theory but also mark a historic milestone in the integration of artificial intelligence into advanced scientific research. The use of Claude in this context highlights the evolving capabilities of AI systems to assist in tasks previously thought to require exclusively human intuition and creativity. As the mathematical community continues to scrutinize the 17-million-line proof, the outcome could reshape how complex mathematical problems are approached in the future.
Key takeaways
- Levent Alpöge claims to have proved the Riemann Hypothesis using Anthropic's Claude AI.
- The proof is a 17-million-line Lean code file that has passed initial verification.
- MIT's Andrew Sutherland estimates the error probability is near zero after independent checks.
- Experts like Kevin Buzzard urge caution, noting the need for human confirmation of the problem statement.
- A valid proof would solve a Millennium Prize Problem worth $1 million.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.