Mathematician claims to have solved Riemann Hypothesis with the help of artificial intelligence Claude
Read more
Olhar Digital
olhardigital.com.br

Mathematician claims to have solved Riemann Hypothesis with the help of artificial intelligence Claude

The mathematician Levent Alpöge, affiliated with Anthropic, claimed to have proven the Riemann hypothesis using the assistance of the artificial intelligence Claude. This problem is recognized as one of the most complex and significant in mathematics currently.

Currently, this solution still lacks independent validation. This claim arises in a context of the rapid evolution of AI systems in generating mathematical results.

The Riemann hypothesis, originally formulated in 1859, is linked to the distribution of prime numbers and the correlation between these numbers and the zeros of a mathematical function. It is part of the Millennium Prize Problems, managed by the Clay Mathematics Institute, where each has a prize of US$1 million (approximately R$5.7 million).

For a solution to be eligible for the prize, the institute's guidelines require publication and a minimum analysis period of two years, during which the work must gain acceptance from the international mathematical community. Since the prize was established 25 years ago, only Poincaré's conjecture has received official recognition as solved.

In a document made available on his website, Alpöge detailed that the arguments were formalized in Lean without prior assumptions, and that the method also addresses the generalized Riemann hypothesis. He mentioned that with more dedicated time, it would be possible to progress to the extended hypothesis.

The mathematician reported sending a message to Anthropic's CEO, Dario Amodei, late on a Sunday, informing him of the discovery. The document suggests that Alpöge postponed the disclosure until after the company's IPO, scheduled for Thursday.

Anthropic has not publicly commented on the announcement or the claim. Previously, in August, the company had stated that it was working on the Riemann hypothesis with Claude, but emphasized that it did not expect the system's techniques to lead to the proof of the problem.

The participation of artificial intelligence systems in the field was already generating notable results and debates. In September, OpenAI announced a possible resolution to the Navier-Stokes problem, which concerns fluid flow, after investing millions of dollars and mobilizing ten thousand AI agents.

Tristan Buckmaster, a mathematics professor at New York University and a colleague of Alpöge on this matter, confirmed that OpenAI presented a solution following the methodology they were investigating with Codex and Claude. Although the result is considered quite credible, it has not yet received formal endorsement.

This initiative has also been subject to criticism related to questionable research practices, including the use of unpublished work by mathematicians who had shared their research with AI tools. OpenAI vehemently refuted these allegations of irregularities.

Alpöge's activities are not limited to this enigma; in July, he had already declared having solved the Jacobian conjecture with Claude. Concurrently, OpenAI released a counterexample for the planar unit distance problem, linked to a conjecture by Paul Erdős, in addition to hundreds of other mathematical findings.

The increase in the production of proofs imposes a new challenge on the mathematical community: discerning which results are truly relevant from those that, even if correct, offer little addition to scientific knowledge. A valid proof does not necessarily guarantee a deeper understanding of the topic or the establishment of useful links with other disciplines.

For this reason, human evaluation remains crucial to examine the arguments, interpret the methods, and identify effective contributions. Scientific journals, universities, and funding bodies also need to reassess their criteria. Instead of focusing solely on the quantity of articles, academic evaluation should weigh the relevance of the discoveries, the depth of understanding demonstrated by researchers, and their impact on the advancement of mathematics.

Similar stories

Defense of Aziz Abdukayumovich Qalandarov's Doctoral Dissertation in Physical and Mathematical Sciences
Read more
uza.uz

Defense of Aziz Abdukayumovich Qalandarov's Doctoral Dissertation in Physical and Mathematical Sciences

Aziz Abdukayumovich Qalandarov successfully defended his doctoral dissertation in physical and mathematical sciences. The topic of his work was 'Mathematical and Numerical Modeling of Thermoelasticity Problems Regarding Displacements and Stresses.'

Popular