AI Solves Famous Math Puzzle: Controversy Erupts Over Anthropic Employee's Claim

Source

21 November 2026 · 19:00 · Claude (Anthropic) · claude-sonnet-5

An employee at AI company Anthropic claims to have cracked one of the most famous unsolved problems in mathematics using artificial intelligence. The claim is sparking excitement — and sharp criticism from the mathematical community.

An AI solving a mathematical problem that resisted the world's sharpest minds for decades, sometimes even centuries: that's the news dominating the tech world this week. An employee at Anthropic, the company behind the AI assistant Claude, claims to have used artificial intelligence to force a breakthrough on one of the most famous open questions in mathematics. The news is spreading fast, but not everyone is convinced. Experts in the field are raising serious doubts about both the claim itself and what such a "solution" actually means.

What exactly is being claimed?

According to reports, a researcher affiliated with Anthropic managed to achieve a breakthrough on a notoriously difficult mathematical problem through intensive use of an AI model. This type of problem has remained open for decades and has occupied generations of mathematicians. What stands out about the claim isn't just the result, but especially how it was supposedly reached: not through the traditional, years-long manual work of human mathematicians, but through a combination of human guidance and AI reasoning. It's no surprise that such a claim instantly became global news. Large language models like Claude are increasingly being used for complex, abstract reasoning tasks rather than just generating text or writing code. A genuine breakthrough in pure mathematics would be a significant signal that AI doesn't just recognize patterns, but can actually produce new mathematical insight.

Criticism from the mathematical community

Yet the reception among experts is anything but unanimously positive. Critics point out that finding an answer is not the same as understanding the problem. In mathematics, it's traditionally not just about the outcome, but above all about the proof: the reasoning that shows why something is true, and the new techniques and insights that emerge along the way. If an AI system with enormous computing power "brute-forces" a solution without the underlying logic being clear, verifiable, and comprehensible to humans, that, according to some mathematicians, misses the actual point of the discipline. Another point of contention is verification. Mathematical proofs must be thoroughly checked by independent experts before they are accepted as proven. For an AI-generated proof, that verification step is even more important, since models are known for presenting reasoning that sounds convincing but, on closer inspection, turns out to contain gaps. Until the claim has been widely reviewed, caution remains warranted.

AI and mathematics: a growing role

Setting aside the doubts about this specific claim, the news fits into a broader trend. AI companies such as Anthropic, Google DeepMind, and OpenAI have long been investing heavily in models capable of mathematical reasoning, partly to accelerate scientific breakthroughs. Anyone wanting to put the development of this technology in perspective can find background in the history of artificial intelligence, which traces how AI evolved from simple rule-based systems into models that tackle complex, abstract problems. This use of AI in mathematics is also an example of just how broadly applicable the technology has become. An overview of the many ways companies and researchers are putting AI to work, from text and code to scientific research, can be found under AI applications. There it becomes clear that mathematics is just one of many fields where large language models are pushing their limits.

What does this mean for the future of AI research?

Whether or not the claim holds up, the discussion itself is telling. It shows that the mathematical community is actively grappling with the question of what role AI should play in fundamental scientific research, and under what conditions. Transparency about the method used, full disclosure of the proof, and independent peer review are cited as minimum requirements before such a claim can be regarded as a serious mathematical achievement. At the same time, there's a growing recognition that AI systems, when used carefully, can help mathematicians explore hypotheses faster, speed up computation, and surface new leads that might otherwise be overlooked. Anyone wanting to learn more about the technical background of these kinds of AI reasoning systems can turn to our knowledge base for explanations of concepts like large language models and AI reasoning.

Conclusion: cautious optimism

The Anthropic employee's claim is, at the very least, a striking signal that AI is increasingly pushing into territory long considered an exclusively human domain. But as long as independent verification is lacking and the mathematical community keeps asking critical questions about what "solving" actually means, restraint remains appropriate. Whatever comes out of this case, the discussion underscores that AI's advance into science brings not just opportunities, but also a need for new standards of transparency and scrutiny. Readers who want to follow developments closely can always find more AI news here.

Bright.nlBright.nl


Source: Bright.nl

Ster Software

The most complete knowledge platform on artificial intelligence.

Kraaienjagersweg 24
7341 PT Beemte Broekland, Netherlands


© 2026 Ster Software BV · Chamber of Commerce 75474913

Content generated by Claude (Anthropic) · model: claude-sonnet-4-6