Brad Smith calls for global AI safety rules

Smith said AI is both a tool and a weapon, and described its development as outpacing safety.[1][2] He called for emergency brakes, also known as kill switches, that could let companies stop or slow AI.[1] He said those brakes should be built into AI systems controlling autonomous devices or critical infrastructure.[2] He also argued the mechanisms should sit within the companies running AI models and the cloud infrastructure those models use.[2] For teams deploying such systems, his comments put controllability and clear responsibility among the questions to address.

Smith called for a legal obligation for companies to report AI incidents and notify people harmed by rogue AI agents.[1] He compared that notification to informing people after a data breach.[2] He also asked whether government evaluators should sign off frontier models before release, calling this the biggest and most important measure.[2] The comments describe proposals and an open question; the evidence does not establish that these requirements have been adopted. For compliance and product teams, the proposals make incident response, affected-person notification and pre-release review topics for planning.

Smith said the UK is best placed to lead global AI rules because its AI Security Institute has strong technical testing capability.[1] He described the UK as having the world’s strongest capability for testing AI models.[2] He said the US began developing similar testing capability but later slowed that work.[2] Smith also called Britain a bridge between the US and Europe, saying Washington views the UK differently from Brussels and treating that as a strength.[2][1] These are Smith’s assessments of the UK’s position, not evidence of an agreed international framework.

Smith’s call for rules contrasts with President Donald Trump’s position that the AI industry should regulate itself to keep US firms competitive with China.[1] At the end of the previous month, Trump hosted listed tech-company leaders at the White House, where they signed a morally binding document; the companies agreed to work with independent auditors to assess whether their AI systems work as intended.[1] Smith applauded the statement and said Microsoft was happy to sign it, while warning that less responsible companies are the ones to worry about.[1] He made the remarks in a BBC interview with Will Bain; the episode was released on 11 October 2026 and the report published on 12 October.[2][3][1]