Microsoft's Brad Smith says AI faces fewer safety rules than milk

Microsoft Vice Chair and President Brad Smith has argued that a carton of milk on a supermarket shelf or a washing machine at home is subject to more safety regulation than artificial intelligence. Speaking on the BBC's "Big Boss Interview" podcast, in remarks reported on October 12, 2026, Smith called for swift rule-making on a global scale.

Smith described AI as both a tool and a weapon, and said governments should move without delay to address the risk of the technology slipping beyond human control. In his view, ordinary consumer goods pass through rigorous safety checks while AI systems lack a comparable framework.

Emergency brakes and independent testing

At the heart of his proposals are so-called kill switches. Smith said companies should be required to have emergency brakes that allow an AI system that has gone out of control to be halted or slowed down.

He also proposed a legal duty to notify affected parties whenever a malicious or rogue AI agent breaches their systems. A further measure would require new AI models to undergo thorough evaluation by independent government inspectors before they reach the market.

Smith pointed out that such oversight has worked well for other consumer products for many years, and said that experience should now be adapted to AI.

Pointing to industry support

Smith said leading developers such as Anthropic, OpenAI and Google DeepMind have also shown support for the idea of controlled progress, which he argued is a reason to begin the regulatory process immediately. He did not detail how those companies view the specific rules he outlined.

Microsoft is one of OpenAI's biggest partners and a central player in the AI race, which makes a call for tougher oversight from its president notable. The comments add to an ongoing international debate over how, and how quickly, governments should set binding safety requirements for advanced AI.