Microsoft president says milk faces more regulation than AI

Microsoft president Brad Smith has argued that AI requires urgent global regulation, warning that it currently faces fewer safety requirements than milk. Speaking on the BBC, Smith said a set of global safety rules should be implemented, including informing victims of system breaches by rogue AI agents and providing an emergency brake.

He suggested that companies should be legally required to submit incident reports, a practice that is common for other products. Smith pointed out that, currently, tech companies such as Anthropic and OpenAI have started disclosing undesirable behavior by AI agents, highlighting the growing debate around the fast-advancing technology.

Some have called for AI development to slow down, sparking a debate about potential motives behind such calls. Smith highlighted Britain as the country best positioned to lead AI regulation, citing its AI Security Institute launched under former Prime Minister Rishi Sunak. The institute receives pre-deployment access to major AI models, allowing it to assess their capabilities and risks.

British officials are considering regulating advanced AI models if current measures prove insufficient.