42 Royal Society Fellows Warn AI Existential Risk Is “Real and Urgent”

Forty-two mathematical fellows of the Royal Society are urging the British Academy of Sciences to alert government and media about existential risks from advanced AI. The open letter includes Fields Medal winners Martin Hairer, Peter Scholze, and Wendelin Werner, and The Decoder reports that none of the signatories are affiliated with AI companies. Their central concern is timing: if the risks become obvious to the wider public only after capabilities accelerate further, meaningful action may come too late.

The letter points to recent progress by leading models from OpenAI and Anthropic, which The Decoder says have solved open research problems within three months. The article says these systems now perform “at the level of the top human mathematicians in many parts of the subject,” while the letter warns of a “significant chance” of superhuman abilities emerging within a similarly short timeframe. For readers, the takeaway is that rapid gains in a technical field like mathematics are being treated as a signal for broader capability growth.

The signatories say the impact on mathematics is not the main reason for the letter. They warn that AI capability growth may advance similarly in areas such as cybersecurity, autonomous weapon control, development of biological and chemical agents, and targeted misinformation. The practical implication is that policymakers and institutions may need to assess frontier AI not just by current use cases, but by how quickly capabilities could transfer into high-risk domains.

The article frames the letter as a push for public and government attention rather than a technical dispute within mathematics. Key signals to watch include how scientific institutions respond, whether governments treat AI lab risk estimates as credible, and whether safeguards keep pace with model capabilities. The most important takeaway: the mathematicians are asking decision-makers to act before risk becomes obvious, not after.