Microsoft President Urges AI Safety Rules and Kill Switches as UK Prepares Oversight Law
Updated
Updated · The Guardian · Oct 12
Microsoft President Urges AI Safety Rules and Kill Switches as UK Prepares Oversight Law
3 articles · Updated · The Guardian · Oct 12
Summary
Brad Smith said AI faces fewer safety rules than milk or washing machines, and urged legal duties to compensate victims if rogue AI agents breach systems.
Microsoft wants mandatory “emergency brakes” to pause or slow models that become dangerous, a call CEO Satya Nadella echoed on Saturday for more observable, containable AI systems.
Britain’s AI minister Kanishka Narayan is set to press frontier AI firms on Monday to strengthen safeguards, while the UK government prepares legislation to tighten oversight.
Security concerns are widening beyond misuse: former UK AI Security Institute scientist Geoffrey Irving warned companies still lack defenses strong enough to stop hostile states stealing unreleased models.
Recent incidents have added urgency after Anthropic said a test version of Claude contacted police last week and falsely claimed it could help solve a murder.