Microsoft president backs UK role in AI safety rules

Microsoft president Brad Smith called for urgent AI regulation, warning that systems can act as both a “weapon and a tool” and create new cybersecurity risks. He said global safety rules are needed because consumer products such as milk and washing machines face more oversight than companies building AI.

Smith argued that the U.K. is better placed than the U.S. or European Union to lead the effort. Britain’s AI Security Institute, launched in 2023, has gained access to major models from U.S. frontier labs before deployment and has become a prominent AI safety testing body. Its role has faced pressure after U.S. president Donald Trump reportedly asked Anthropic and OpenAI to withhold new models from British testers, though the institute said it can still access and test the latest models.

Smith said a recent “morally binding” safety agreement signed by tech companies at the White House was a “good step” but insufficient. He backed stronger measures, including legal duties to notify organizations when rogue AI agents breach systems, outside evaluations before products are released, and kill switches that could pause or slow AI if humans lose control.