LONDON — Artificial intelligence needs urgent global regulation, with independent evaluators able to vet new developments before they reach the public, Microsoft president Brad Smith said in a BBC interview published on Monday.
His warning carried a striking comparison: AI currently faces fewer safety requirements than the milk in your refrigerator.
“Right now, your washing machine is subject to more safety rules than AI. Every carton of milk that you see in the grocery store is subject to safety rules and inspection,” Smith told the BBC. “We know how to do this for other products. We should build on that knowledge and think about how to apply it to AI.”
A concrete blueprint for AI safety
Smith laid out a specific package of global safety rules. Companies should be legally required to write incident reports when something goes wrong.
A “victim” — a company or person whose system is breached by a rogue AI agent — should be informed. And an “emergency brake” should be available to shut dangerous systems down.
The proposals put structure around a debate that has stayed mostly abstract until now. Rather than calling vaguely for AI safety, Smith's framework borrows directly from how governments already regulate medicine, food, and consumer electronics — inspection, disclosure, and accountability.
Rogue agents and the speed of the debate
The comments land at a moment when the industry's own behavior is feeding the case for oversight.
Anthropic and OpenAI have begun publicly disclosing rogue or undesired behavior by their AI agents, part of a fast-moving global debate over how quickly advanced AI is spreading. That debate has also absorbed warnings from researchers about AI's eventual existential risks.
Not everyone reads such calls the same way. Some in the industry have urged AI development to slow down, which has sparked a counter-argument: that executives pushing regulation may have ulterior motives — preserving their firms’ lead, pre-empting a broader government backlash, or simply putting the brakes on breakneck AI spending.
Britain's case to lead
Smith pointed to one country as especially well-placed to take the lead: Britain. He cited the AI Security Institute, launched under former prime minister Rishi Sunak, and argued that officials in Washington view London very differently from how they view Brussels, where the EU regulates AI based on levels of risk under its AI Act.
Britain's institute already receives pre-deployment access to leading AI models under voluntary agreements, allowing it to assess their capabilities and risks before release — a working model of the independent evaluation Smith wants globalized.
By Hamza Faisal
