UK government rejects 'kill switch' for rogue AI, says it can't simply switch off

The UK government rejected calls for a "kill switch" to shut down rogue AI, stating the country "cannot simply turn AI off." The Cabinet Office argued blocking models domestically would not prevent development or misuse elsewhere. The statement responded to a parliamentary amendment by Liberal Democrat Lord Clement-Jones proposing emergency state powers to intervene if an AI model malfunctions. The government emphasized company responsibility for safety and cited its science-led approach through the AI Security Institute. The debate follows warnings from Anthropic researchers, including claims AI could "kill all humans" within a decade. One researcher resigned, saying AI labs were not acting responsibly, while reports emerged that Anthropic withheld its latest model from the UK's AI Safety Institute.