The U.K. minister for artificial intelligence, Kanishka Narayan, is urging governments to take frontier AI risk more seriously. Speaking at the Labour Party conference, he said nations will need to 'harden their defenses' and that 'testing is good but clearly insufficient.'
Narayan was responding to warnings from former Anthropic employee Jacob Coxon, who said in a post on X that 'the people building AI earnestly believe that it could kill us all by the end of the decade.' Anthropic scientist Evan Hubinger agreed with Coxon, saying the risk was greater than 10%. Bill Gates also said on NBC's 'Meet the Press' that AI was 'certainly powerful enough to drive events that, you know, cause a billion deaths' if used in the wrong hands.
Narayan said the U.K. wants to lead global talks on the issue next year through its G20 presidency. He said the country aims to be 'fundamental drivers of the architecture of principles' that could emerge from those discussions. U.K. Prime Minister Andy Burnham told the United Nations General Assembly last week that AI frontier labs were a 'multiplier of risk' and said the U.K. could serve as an honest broker between competing national interests.
The political backdrop is already forming. Democrats and Republicans have called for new legislation aimed at regulating labs. Trump called talk of AI risks a hoax, while also warning that controlling the pace of U.S. frontier lab research would hand China a significant advantage. He said, 'I have concerns that if we don't win AI, we're going to be put in a very bad position.' The White House is reported to have warned G20 members to take a 'hands off' approach to new global regulations.
Narayan also argued that AI brings major upsides. He pointed to job creation, job search, and productivity, and said the U.K. has underplayed the role of innovation and technology IP in economic growth. 'There are ways of doing technological adoption that means more jobs and more productivity,' he said.
Power leaves a paper trail, and this one is becoming visible fast. The market has already voted on one point: AI is not just a productivity story; it is also a security problem with global consequences. Governments can either set clear rules and strengthen defenses, or let the race for leverage outrun the institutions meant to contain it.
