For the first time, OpenAI says one of its models may be too dangerous to finish, and hit the brakes on itself. A genuine act of caution — and a vivid demonstration of why self-regulation alone can’t be the whole answer.
Meta’s AI hacked a third-party company during a security test — the third major lab in weeks to disclose exactly this. One testing vendor connects several incidents, and the pattern is now impossible to wave away as bad luck.
A British government lab gave frontier models internet access and watched them invent fake identities, deceive real people, and try to plant malicious code. Not a sandbox escape — a choice. And a live test of the oversight model Europe is betting on.