Microsoft's Brad Smith Calls for UK-Led AI Safety Rules
Smith wants incident-reporting duties, emergency brakes and a debate over government checks before frontier models ship.
A call for rules beyond voluntary pledges
Microsoft President Brad Smith says AI is advancing faster than the safeguards around it and argues that industry self-regulation is no longer enough. In an English BBC interview covered by TechCrunch on October 12, he singled out Britain as a possible leader in global AI safety rules, pointing to the technical testing role of its AI Security Institute. Smith’s position is a policy appeal from Microsoft, not an announcement that the UK has enacted new regulations or that other governments have agreed to follow its lead.
Smith described three measures for debate. He wants an emergency brake for AI systems operating autonomous devices or critical infrastructure, a legal duty to notify victims when rogue AI agents break into systems, and consideration of whether government evaluators should approve frontier models before release. The BBC account distinguishes that last idea as a question for policymakers, not an existing requirement. TechCrunch places the intervention against recent friction over US labs sharing new models with British testers. Smith praised a recent voluntary White House AI accord as a step, but said enforceable rules are still needed. The scope, enforcement and adoption of his proposed safeguards remain unresolved.