AI is less regulated than milk, Microsoft president says

Brad Smith, the president of Microsoft, says that artificial intelligence — one of the most powerful technologies in the world — is less regulated than a carton of milk. Speaking to the BBC, Smith argued that AI urgently needs global safety rules to protect people from harm. He made the comments as debate grows around the world about how to control AI before it becomes too dangerous to manage.

Smith said that right now, washing machines and milk cartons must follow strict safety rules, but AI does not. He used these everyday examples to show how far behind the rules are for such a powerful technology. 'Every carton of milk that you see in the grocery store is subject to safety rules and inspection,' he said. He believes the same kind of care should be applied to AI.

Smith called AI both a 'tool and a weapon.' He said companies need 'emergency brakes' — sometimes called 'kill switches' — that could pause or slow down AI if it started to act in dangerous ways. He also said companies should be required to write 'incident reports' to explain what went wrong when AI causes harm. On top of that, he said victims should be told by law if an AI system is used against them.

Smith's views put him at odds with U.S. President Donald Trump. Trump believes the AI industry should regulate itself so that American companies don't fall behind China. Trump has repeatedly said that 'whoever wins AI, wins.' He has pushed back against any efforts to slow down AI development.

Smith does not think the U.S. government would be totally against having the United Kingdom lead the effort to create global AI rules. He said the UK has a special advantage because of its AI Security Institute, a government body set up to study AI safety. 'Most people that I know would say, name the country that is the leader in having this technical capability, they would say it's the United Kingdom,' Smith told the BBC.

Smith said the UK is better placed than the European Union to set these rules, partly because Washington, D.C., views the UK more favorably than it views Brussels. He pointed to risks like AI agents that have 'escaped human control' and AI models powerful enough to threaten the world's cybersecurity. 'The moment has arrived,' he said, calling for calm but firm action to bring countries and companies together.

At the end of last month, President Trump met with the leaders of major tech companies — including OpenAI, Google, Nvidia, and Meta — at the White House. The executives signed a document described as 'morally binding,' agreeing to work with independent auditors to make sure their AI systems work as intended. Smith said Microsoft is 'happy to sign it' and called it a good step, but said it is no longer enough on its own.

Smith warned that voluntary agreements only work when all companies play by the rules. 'The good companies will do what you want,' he said. 'It's the companies that are not as responsible that you need to worry about.' He believes that without official laws, irresponsible companies could still develop AI in dangerous ways without any consequences.