42 mathematicians warn that advanced AI poses urgent existential risks
Forty-two Fellows of the Royal Society have signed an open letter urging the president of the institution to ask the British Academy of Sciences to alert the government and media about existential risks from advanced artificial intelligence. The signatories include Fields Medal winners Martin Hairer, Peter Scholze and Wendelin Werner. None is affiliated with an AI company.
The letter says leading models from OpenAI and Anthropic solved open research problems within three months, including one of the seven Millennium Problems. It adds that a second may soon be solved. According to the mathematicians, the systems already perform at the level of leading human mathematicians in many areas and could develop superhuman abilities within a similarly short period.
The concern extends beyond mathematics. The signatories say capability gains could also accelerate cybersecurity threats, autonomous weapons control, the creation of biological and chemical agents, and targeted misinformation. They argue that estimates from AI laboratories placing existential risk above 10 percent should not be dismissed as hype, warning that public recognition of the danger could come too late for action.
Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.