AI is less regulated than milk, says Microsoft president

Microsoft president Brad Smith warns that artificial intelligence is currently subject to fewer safety rules than everyday products like milk cartons and washing machines, and he urges immediate global regulation. He frames AI as both a powerful tool and a potential weapon, citing risks such as rogue AI agents and severe cybersecurity threats. Smith proposes concrete measures: legal obligations for firms to file incident reports and notify victims if systems are breached, built-in "kill switches" to pause dangerous behavior, and independent government evaluators to vet systems before release. He contrasts this with calls from some leaders for industry self-regulation, arguing that voluntary regimes will fail to restrain less responsible actors.

Smith applauds a recent industry commitment to independent auditing but says voluntary pledges are not enough and calls for legally enforceable standards. He recommends the United Kingdom as a natural coordinator of international rules because of its AI Security Institute and technical strengths, and urges calm yet decisive collaboration to build safety frameworks now. His core argument is that enforceable reporting, oversight, and emergency controls are practical extensions of existing product-safety approaches and must be implemented immediately to manage AI’s systemic risks.