British lawmakers want a legal off switch for AI that gets too powerful.
Lord Tim Clement-Jones has proposed an amendment to the Cyber Security and Resilience Bill, now moving through Parliament, that would give the UK a "democratically accountable" way to halt a runaway AI system before it compromises critical infrastructure. He's framed it as a last-resort safety net, not a routine control. Separately, Labour MP Alex Sobel plans to introduce an AI Security Bill this month that would go further, effectively halting development of superintelligent AI and making the UK the first G7 country to do so. That bill has backing from ControlAI, a UK advocacy group led by Andrea Miotti, who previously worked at the AI safety company Conjecture.
This matters because it's a shift from AI companies self-policing to lawmakers writing enforcement mechanisms into statute. The UK effort isn't isolated either - US legislators are drafting a similar AI Kill Switch Act, though it's still early. If either UK bill passes, it would give regulators explicit legal authority to intervene, not just guidance to ignore.
There's some irony here: AI labs spent years marketing certain models as "too dangerous" for general release, and now legislators are taking that framing at face value and writing kill switches into law.