Your grocery store’s dairy aisle might seem like an unlikely place to start a conversation about AI safety. But Microsoft President Brad Smith says it offers a useful comparison for how the technology should be regulated.
Speaking to the BBC’s Big Boss Interview podcast on Oct. 11, Smith called for global AI safety rules, arguing that everyday products like milk face more safeguards than increasingly powerful AI systems.
“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.
His proposed protections go beyond asking companies to promise their technology is safe. Smith wants legal requirements for companies to report safety incidents and notify victims when rogue AI agents breach their systems. He also called for “emergency brakes” that could pause or slow AI systems, alongside rules determining when independent evaluators should review new technology before its release.
“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.”
Smith’s comments echo recent warnings from Bill Gates. The push for stronger AI protections also comes as President Donald Trump has favored industry self-regulation. At a White House meeting in September, major AI companies agreed to voluntary safety standards, including working with independent auditors to assess whether their systems behave as intended.
Smith welcomed that agreement and said Microsoft was happy to sign it. But he argued that voluntary commitments would no longer be enough. “It’s a good step, it’s a necessary step… But I don’t think it’s sufficient any longer,” he told the BBC.
The problem, in his view, is that responsible companies will follow voluntary rules. Less responsible ones may not.
Who gets to set the rules?
That question has also surfaced in Mashable’s reporting on AI leaders’ calls for stronger safety standards. As Rebecca Ruiz reported in September, experts caution that companies can support safeguards while still seeking control over how those safeguards work.
Pete Furlong, senior manager of policy and research at the Center for Humane Technology, told Mashable that companies’ transparency efforts need to come with expectations beyond those they set themselves.
“It can’t be entirely up to these companies to dictate the terms of their transparency,” he said.
Smith suggested the UK could help lead the development of global rules, pointing to the technical expertise of its AI Security Institute. The institute already gets access to major AI models before deployment through voluntary agreements, allowing it to assess their capabilities and risks.
For Smith, the next step is making safety obligations enforceable — including requiring companies to tell people when something goes wrong.
“It’s the companies that are not as responsible that you need to worry about,” he said.
Want to learn more about getting the best out of your tech? Sign up for Mashable’s Top Stories and Deals newsletters today.