Brad Smith urged safety rules for AI, including independent evaluations, incident reporting, and mechanisms to halt systems behaving dangerously.
Microsoft President Brad Smith has called for stronger safety rules for artificial intelligence, arguing in a BBC interview that AI faces less formal safety oversight than everyday products such as milk and washing machines. The Guardian reported his remarks on October 12.
Smith said AI can function as both a useful tool and a weapon, and urged companies and policymakers to establish safeguards. Among the measures he supported were “emergency brakes” or kill switches to pause or slow systems if they behave dangerously, independent assessments before release, and a legal obligation to notify people whose systems are breached by rogue AI agents. These proposals address both pre-release testing and what happens when AI tools are deployed or compromised.
The comments add to a wider debate about how responsibility for AI risks should be divided among technology firms, regulators, and users. They also follow calls from Microsoft CEO Satya Nadella for containment mechanisms that make powerful systems’ behavior observable and their actions controllable. Smith’s comments are proposals, not an announcement of new Microsoft products or binding regulation.
For companies deploying AI agents, incident reporting and reliable shutdown mechanisms could become important governance and operational questions if such ideas are adopted by lawmakers or industry standards bodies. Independent evaluations could also affect how developers test and document systems before launch. Smith did not set out a specific legislative text or implementation timetable in the interview. The significance of the remarks is that a senior Microsoft leader is publicly advocating additional safeguards, including obligations that could apply beyond the company itself, at a time of heightened scrutiny over agent behavior and AI oversight.