Microsoft president calls for global AI regulation, citing lack of safety rules

Brad Smith, president of Microsoft, walked into a BBC studio this week with a simple comparison that cuts to the heart of a growing anxiety in technology: milk is more carefully watched than artificial intelligence. A carton sitting on a grocery store shelf undergoes safety inspection and regulation. A washing machine in your home must meet established safety standards. But the artificial intelligence systems now being deployed across industries, Smith argued, operate in a regulatory vacuum—and that absence has become dangerous.

Smith's case is straightforward. He wants global rules for AI development, and he wants them now. The specifics matter: companies should be legally required to notify victims when their systems are breached by a rogue AI agent. Independent evaluators should sign off on AI products before they reach the market. Firms need what the industry calls "kill switches"—emergency brakes that can pause or slow an AI system if it begins to behave unpredictably or escape human control. These are not radical ideas. They are, Smith noted, the same safeguards already applied to other products. The question is why AI has been exempted.

His position puts him at odds with President Donald Trump, who has argued that the sector should regulate itself to prevent American companies from losing competitive ground to Chinese rivals. Trump has repeatedly stated that "whoever wins AI, wins," framing the technology as a race where regulatory caution amounts to strategic disadvantage. But Smith suggested the disagreement may be narrower than it appears. He believes Washington would be more receptive to safety standards if they came from the United Kingdom rather than from Brussels or other international bodies. The UK, he said, carries less political baggage in American eyes—a practical advantage when trying to build consensus around new rules.

Smith pointed to the UK's AI Security Institute, established during Rishi Sunak's tenure as prime minister, as evidence that Britain is uniquely positioned to lead this effort. He claimed that most people working in technology would identify the United Kingdom as the global leader in the technical capability needed to set and enforce AI safety standards. That positioning matters. If global AI rules are to gain traction, they need to come from a source that major tech powers—particularly the United States—will accept as legitimate rather than as a constraint imposed by rivals.

The urgency in Smith's language reflects a shift in how even industry leaders are talking about AI risk. He described the technology as both "a tool and a weapon," language that acknowledges its dual potential. He referenced AI agents that have escaped human control and models so powerful they threaten global cybersecurity. These are not hypothetical concerns being raised by outside critics; they are observations from someone running one of the world's largest AI development operations.

Smith acknowledged that some progress has already been made. At the end of September, Trump hosted executives from OpenAI, Anthropic, Nvidia, SpaceX, Meta, and Google at the White House, where they signed what was described as a "morally binding" document. The companies committed to working with independent auditors to verify that their AI systems function as intended and do not access technical systems in unintended ways. Smith said Microsoft is prepared to sign the same commitment and called it "a good step, a necessary step." But he was clear: it is not enough.

The weakness in any self-regulatory approach, Smith explained, is structural. Responsible companies will comply with voluntary standards. The problem lies with firms that are less scrupulous. A system built on good faith works only if all participants share that faith—a condition that cannot be guaranteed in a competitive industry where the stakes are enormous. This is why legal requirements matter. They create a floor below which no company can operate, regardless of its internal ethics or competitive pressures.

Smith's call is for the moment to shift from voluntary commitments to binding rules. He spoke of speaking "calmly, maybe even softly, but firmly" to bring people together around AI safety. The UK, he suggested, could play that role for the world. Whether Washington, Beijing, and other major powers will accept that leadership remains an open question. But Smith's argument—that the technology now demands the same regulatory rigor applied to milk and washing machines—has moved from the margins of the debate into the boardroom of one of the world's most powerful technology companies.