Microsoft's Brad Smith says AI needs rules now — and the UK should write them

Microsoft President Brad Smith says consumer products like milk and washing machines face more regulatory oversight than the companies building artificial intelligence — and he wants that to change, with the United Kingdom leading the effort.

Speaking on a recent BBC podcast, Smith called AI both a "weapon and a tool" that is putting the world's cybersecurity at risk, and said the "moment has arrived" to institute global safety rules.

The intervention matters because it comes from the president of one of the world's three largest AI investors, at a moment when voluntary commitments are the only guardrails around frontier model development in the United States. It also lands amid open friction between Washington and London over who gets to test American AI systems before the public does.

Why does Smith think the UK should lead?

Smith argued that the U.K., and not the U.S. or European Union, is best placed to lead on global AI safety rules. The core of his case is institutional. Britain's AI Security Institute, launched in 2023, receives access to major models from U.S. frontier labs before they are deployed and has quickly become a leader in AI safety testing.

No equivalent pre-deployment testing relationship exists in the United States, and the EU's regulatory approach has made Brussels a target for American political resistance.

"In Washington, D.C., people don't look at the U.K. and they don't look at London the way they look at Brussels," Smith told the BBC. "That's a strength."

The subtext is blunt: the UK is close enough to the U.S. to be trusted with pre-deployment access to frontier models, but foreign enough to act as a neutral safety tester. Smith believes that positioning lets the UK regulate AI in ways the U.S. cannot.

What has Trump done about British AI testing?

The UK's testing role has been under threat in recent weeks. U.S. President Donald Trump has warned against "globalist" organizations, which he claims are interfering with American companies. Last month, he reportedly asked Anthropic and OpenAI to withhold new models from British testers.

The British institute pushed back on that pressure this week. AISI confirmed on Monday that it is able to access and test the latest models. Smith's public endorsement of the UK's role, delivered days later, adds corporate weight to the institute's position at precisely the moment the White House is trying to curtail it.

The episode illustrates the stakes of the story. Pre-deployment access is the only mechanism that currently gives any outside body a look at frontier AI systems before millions of users touch them. If that access breaks down, safety testing shifts entirely to the labs themselves.

What rules does Smith want?

Smith outlined several AI safety measures he would like to see put in place:

• A legal obligation to alert an organization if its system is breached by rogue AI agents.
• Approval from outside evaluators before an AI product is released.
• Kill-switches that can pause or slow down AI systems if humans lose control of them.

He spoke positively about the kill-switch idea in particular, framing it as a mechanism for the moment when human oversight fails rather than a hypothetical.

None of these rules exists today in binding form for frontier AI developers in the U.S. or the UK. Smith's proposals would convert the current regime of private evaluation agreements into enforceable legal duties.

Why does Smith reject self-regulation?

Smith's comments come amid rallying cries from tech leaders, including Anthropic's Dario Amodei and OpenAI's Sam Altman, to slow the pace of AI development, following several incidents where AI agents went rogue.

Last month, Trump hosted leaders of OpenAI, Anthropic, Nvidia, AMD, SpaceX, Meta and Google at the White House to discuss AI safety. They signed a "morally binding" document in which they agreed to work with "independent auditors" to assess the safety and capabilities of their AI models.

Smith called the document a "good step," but said it stops short of the type of regulation that is necessary. His reasoning focuses on the asymmetry of voluntary compliance. The problem with voluntary self-regulation, he said, is that the "good companies" will comply.

"It's the companies that are not as responsible that you need to worry about," Smith said.

That argument cuts against the current direction of U.S. policy, which relies on industry-led commitments. It also puts Microsoft's president in the awkward position of asking governments to regulate his own industry — while the largest AI developers in the U.S. operate under no binding pre-deployment safety regime at all.

What happens next?

Smith closed with an appeal for coalition-building rather than confrontation. "Let's speak calmly, maybe even softly, but firmly, and start to bring people together," he said. "I hope that's the role we can play as a company. I hope that's the role the U.K. can play as a country."

The immediate test of that framing is already underway. AISI retains access to the latest models as of Monday, but Trump has shown willingness to ask labs to cut British testers off. Whether the UK's pre-deployment testing role survives will depend on whether more companies adopt Smith's public stance — or whether the White House escalates its campaign against the "globalist" organizations it says are interfering with American firms.