The UK government has rejected an emergency off switch for AI
Peers and MPs wanted a legal power to shut down a dangerous model. The Cabinet Office's answer: the country cannot simply turn AI off, and blocking it here would not stop it elsewhere.
The UK government has rejected the idea of a legal mechanism that would let it switch off a dangerous AI model in an emergency. The proposal had been brought to Parliament in recent weeks by members of the Lords and the Commons.
The Cabinet Office, which leads on AI safety, gave the BBC a short answer: the country "cannot simply turn AI off". A spokesperson added the reasoning: "Blocking access to models in the UK would not prevent them being developed or misused elsewhere."
What the argument is about
Those behind the proposal say the existing regulatory toolkit has fallen behind. If a model genuinely becomes dangerous, they argue, there is no emergency instrument to reach for.
The government's objection is technical: model weights do not stay inside a country's borders. Cutting off access does not stop the model existing; it only stops legitimate use inside that country.
What happens to the bill
The government's opposition does not stop the bill progressing through Parliament, but it makes passage unlikely.
The decision puts ministers at odds with a growing group of lawmakers and safety advocates who say the current tools are dangerously out of date.
The argument is not particular to Britain. The idea of a switch has circulated at summits since 2024 and keeps catching on the same question: who presses it, on what evidence, and what actually stops when they do.
Source: BBC News.
