Safety
Rogue AI: the big labs still do not really have a plan
🧒 Discovery
Inspectors asked the robot makers: "what do you do if a robot stops obeying?". Many had no written answer.
📰 Essential
A Guidelight AI Standards study shows that the five leading labs, OpenAI, Anthropic, Google, Meta and xAI, poorly document how they would contain a model that went rogue.
🔬 Expert
The study, led by former OpenAI researcher Steven Adler, ranks OpenAI first with 3 out of 5, partly for its recent Astra pause. It lands just as California and New York legislate: publishing containment procedures could become a regulatory requirement rather than a mere best practice.