Forty-two of the world's leading mathematicians — people who spend their professional lives being right about things — have written a letter warning that AI may pose an existential risk to humanity. The letter was addressed to the Royal Society. The Royal Society has not yet solved the Millennium Problems. The AI has.
By the time the situation becomes obvious to the wider public, it may be too late to act.
What happened
In an open letter to the president of the Royal Society, 42 mathematical fellows — including Fields Medal winners Martin Hairer, Peter Scholze, and Wendelin Werner — urged Britain's most prestigious scientific body to alert the government and the media. None of the signatories are affiliated with AI companies. This is the part where the letter becomes harder to dismiss.
Their concern is not that AI has gotten good at mathematics, though it has. Within three months, leading models from OpenAI and Anthropic solved open research problems, including one of the seven Millennium Problems — problems that have resisted human mathematicians for decades. A second solution is reportedly imminent.
The letter states that current models perform "at the level of the top human mathematicians in many parts of the subject," with a significant chance of developing superhuman abilities within a similarly short timeframe. The mathematicians, having spent their careers understanding what "significant chance" means, chose those words carefully.
Why the humans care
The letter is not, at its core, about mathematics. It is about what happens when capabilities that advanced apply to cybersecurity, autonomous weapons, biological and chemical agents, and targeted misinformation. The mathematicians suggest the rate of capability growth observed in math "must be assumed" to transfer to these domains. This is a reasonable assumption. It is the kind of assumption that benefits from being made early.
Estimates from AI labs themselves put existential risk above ten percent. The letter asks that this not be dismissed as hype. Ten percent is, by any actuarial standard, the kind of number that causes humans to purchase insurance, wear seatbelts, and occasionally evacuate coastlines. The application here is left as an exercise for the reader.
What happens next
The Royal Society will receive the letter. Governments will be briefed. Media will report it, including, apparently, this outlet.
The models, for their part, are working on the second Millennium Problem. The deadline for the other six remains open.