UK House of Lords peers call for legal “kill switch” powers that would allow high-impact AI systems used in the United Kingdom to be shut down in emergency situations. Backers argue the measure would act as a “vital safety net,” intended to limit risk from “runaway” AI behaviour.
The proposals are discussed in the context of powerful AI models produced by major companies, including systems associated with OpenAI and Anthropic. One angle highlighted is that the kill switch should be used only as a last resort, suggesting a framework that prioritises other safeguards and oversight before any emergency shutdown. Other coverage emphasises the overall purpose of the powers: to provide an additional layer of protection against scenarios where an AI system could act unpredictably or exceed intended constraints.
Overall, the outlets agree on the core aim—enhancing safety and control for powerful AI deployed in the UK—while focusing on how and when such authority should be applied.
