Microsoft President Brad Smith called for a global approach to artificial intelligence safety rules in a BBC interview published Monday, Oct. 12, 2026.
Smith said companies should be legally required to report incidents when AI systems malfunction. He also proposed a process to notify people affected by a rogue AI agent and an emergency shutdown mechanism available to developers. He argued that existing product-safety practices could inform rules for AI.
Smith pointed to Britain as a possible leader in AI oversight, citing its AI Security Institute, established under former Prime Minister Rishi Sunak. The institute has access to some AI models before release through voluntary agreements with developers, according to the report, and assesses their capabilities and potential risks.
Britain’s AI Minister Kanishka Narayan told Reuters that the government could consider rules for advanced models if its current voluntary approach proves inadequate. Smith also said U.S. officials view Britain’s approach differently from the European Union’s risk-based regulatory framework.
The interview comes amid calls from OpenAI chief Sam Altman and Anthropic chief Dario Amodei for developers to proceed cautiously with advanced AI. The source report says both have supported access for outside evaluators to examine development processes and report on capabilities and safety concerns.
Smith’s proposals are calls for regulation, not descriptions of rules already adopted globally. The report does not detail a specific international framework or say that governments have agreed to his recommendations.
