Microsoft's Brad Smith says AI has fewer safety rules than a washing machine

Microsoft president Brad Smith has called for binding international safety rules for artificial intelligence, arguing that the technology is now less tightly regulated than everyday household goods. In a BBC interview published over the weekend of October 11 and 12, 2026, Smith said: "Right now, your washing machine is subject to more safety rules than AI." He also pointed out that cartons of milk on supermarket shelves face safety rules and inspection, while advanced AI systems mostly do not. He set out a list of measures he wants governments to adopt, and said the United Kingdom was well placed to lead the effort.

Smith's comments follow remarks by Microsoft chief executive Satya Nadella, reported on October 10, that AI models need an emergency brake. Smith has now turned that idea into a wider list of legal duties. It is unusual for a senior executive at one of the world's largest AI companies to say in public that voluntary commitments are not enough. Smith's argument, as reported, is that responsible firms already follow voluntary expectations, but the greater danger comes from less careful companies that face no legal duty to do the same. Binding rules, in his view, would close that gap.

The proposals are specific. Incident reporting would work much like the reporting systems used in aviation or cybersecurity, where failures are logged so that others can learn from them. Victim notification would mirror data-breach laws that already exist in many countries. An emergency brake addresses a newer worry: AI agents that can take actions on their own, such as moving money or changing computer systems, and that could cause harm before a person steps in.

Smith's choice of Britain as a leader is also notable. He said officials in Washington view London very differently from the way they view Brussels, whose AI Act many US firms see as heavy-handed. Britain's AI Security Institute already tests models before release under voluntary deals. Britain's AI minister, Kanishka Narayan, has said the country would consider regulating advanced models if the current system proved insufficient.

For India's AI developers and the many Indian firms that build on top of models from Microsoft, OpenAI, Google and others, any shared international standard would matter. Rules on testing, incident reporting and kill switches set in one major market tend to shape how global companies build their products everywhere.

Short-term: The remarks add weight to calls for mandatory AI rules at a time when the US government favours self-regulation, and they put pressure on rival AI labs to say where they stand.

Long-term: If governments adopt incident reporting and emergency-stop requirements, AI products, especially autonomous agents, may need new safety features before they can be sold.

Who is affected: AI companies, businesses deploying AI agents, regulators in the UK, US and EU, and users whose data or systems could be touched by AI tools.

Questions and Answers