The UK government rejects proposals for a legal “kill switch” that would allow it to switch off access to dangerous AI models in an emergency. The idea is raised in Parliament by some members as fears grow about potential misuse or risks from advanced AI.

The government’s AI-safety lead, the Cabinet Office, says the UK “cannot simply turn AI off.” Officials argue that blocking access to models inside the UK would not stop the same systems from being developed or misused elsewhere. A spokesperson tells BBC News that limiting UK availability would not address the broader risk.

Outlets also note that the government’s stance does not automatically halt the proposal’s progress through Parliament, though it makes passage less likely. Criticism highlighted in commentary emphasizes that disabling access domestically would not prevent the technology from existing or spreading internationally, with former OpenAI researcher Daniel Kokotajilo referenced on the limited protective value of such a measure.