Microsoft president Brad Smith says AI needs urgent regulation. He points out that everyday goods such as milk and washing machines are subject to more oversight than the companies developing AI.
On a recent BBC podcast, Smith described AI as both a “weapon and a tool” and warned that it puts global cybersecurity at risk. He said the “moment has arrived” for international safety rules, arguing that the U.K. is better placed to lead than the U.S. or the European Union.
Launched in 2023, Britain’s AI Security Institute (AISI) can examine major models from U.S. frontier labs before they are deployed. The institute has quickly become a leader in AI safety testing, but that role has been under threat in recent weeks. U.S. President Donald Trump cautioned against “globalist” organizations, which he says interfere with American companies. Last month, Trump reportedly asked Anthropic and OpenAI not to share new models with British testers. On Monday, however, AISI confirmed it had access to the latest models and could test them.
In Smith’s view, the U.K. can regulate AI where the U.S. cannot. He told the BBC, “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.” He described that as “a strength.”
Calls to slow AI development have come from several tech leaders, including Anthropic’s Dario Amodei and OpenAI’s Sam Altman, following several incidents in which AI agents went rogue. Last month, Trump brought leaders from OpenAI, Anthropic, Nvidia, AMD, SpaceX, Meta and Google to the White House to discuss AI safety. They signed a “morally binding” document pledging to work with “independent auditors” to evaluate the safety and capabilities of their AI models.
Smith described the document as a “good step,” but said it does not go far enough to provide the regulation he believes is needed. He said the problem with voluntary self-regulation is that “good companies” will comply. “It’s the companies that are not as responsible that you need to worry about,” he said.
Smith outlined several AI safety rules he’d like to see put in place, including a legal obligation to alert an organization if its system is breached by rogue AI agents. He also appears to favor AI labs getting approval from outside evaluators before an AI product is released. Smith spoke positively about kill-switches that could pause or slow AI if humans lose control of it.
“Let’s speak calmly, maybe even softly, but firmly, and start to bring people together,” Smith 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.”
