Microsoft’s Brad Smith Warns AI Lags Milk in Safety Rules as Calls for Global Standards Grow

Microsoft President Brad Smith says AI faces fewer safety rules than everyday household goods. “Right now, your washing machine is subject to more safety rules than AI. Every carton of milk that you see in the grocery store is subject to safety rules and inspection,” Smith said on the BBC’s Big Boss Interview podcast, in comments the BBC published on Monday, Oct. 12.

Smith called AI both a “tool and a weapon” and said global safety rules are needed now. “Let’s get started today,” he said.

He laid out several measures. Firms should have a legal obligation to tell a “victim” if their system is breached by a rogue AI agent, and to write “incident reports” when the technology goes rogue. He suggested independent government evaluators should vet new AI products, and said rules should decide when outside evaluators sign off on a product before release. And developers need “emergency brakes,” often called kill switches, to pause or slow down AI if it gets out of control. “We know how to do this for other products,” Smith said. “We should build on that knowledge and think about how to apply it to AI.”

Beyond self-regulation

The US government has favored letting the industry regulate itself, the BBC reported. At the end of September, the leaders of OpenAI, Anthropic, Nvidia, SpaceX, Meta and Google signed a “morally binding” document at the White House agreeing to work with “independent auditors” to check that their AI systems work as intended. Smith said he applauds the statement and Microsoft is “happy to sign it.” But, he told the BBC, “It’s a good step, it’s a necessary step… But I don’t think it’s sufficient any longer.” The challenge for any “self-regulatory regime,” he said, is that “the good companies will do what you want.” He added: “It’s the companies that are not as responsible that you need to worry about.”

Nadella’s emergency brake

Microsoft CEO Satya Nadella made a similar case in a long post on X, a day after Smith first made the milk comparison in an interview with ITV News, The Independent reported. “We must assume a model is compromised and contain it from the start,” Nadella wrote. “Think of it like an emergency brake. Any authorised person should always be able to pause or shut down a model mid-task. More advanced models will require more advanced containment technologies that we need to standardise on.”

The kill-switch idea has gained traction in recent weeks after a string of cyberattacks and other incidents involving frontier AI models, according to The Independent. The heads of Anthropic, OpenAI and Google DeepMind have all backed the idea of slowing AI development, the BBC said.

Why Smith points to the UK

Smith said the UK is best placed to lead on global rules because of its AI Security Institute, launched under former Prime Minister Rishi Sunak as the AI Safety Institute. The institute gets access to major models from US frontier labs before they are deployed, TechCrunch reported. The institute confirmed Monday that it can still access and test the latest models, after reports last month that Anthropic and OpenAI had been asked by the US administration to withhold new models from British testers.

“In Washington DC, people don’t look at the UK and they don’t look at London the way they look at Brussels – that’s a strength,” Smith said. “When you look at AI agents that have escaped human control, when you look at models that are so powerful that they’re putting the world’s cybersecurity at risk. I would say: The moment has arrived.”