12 October 2026

UK minister urges AI labs to halt releases if risks can't be controlled

First reported

The Guardian ran this on 11 October, a day before the other 2 sources picked it up.

Kanishka Narayan, the UK's AI minister, told companies building the most powerful AI systems to heed scientists' warnings and withhold releases if risks cannot be controlled.
The UK is strengthening its cyber defences, including exploring new legislation to protect essential services like hospitals and utilities from AI-related risks.