Robotics & Physical AIJul 2, 2026

Anthropic + Glasswing partners propose an industry jailbreak-severity score

After the Fable 5 recall, Anthropic and Glasswing partners (Amazon, Microsoft, Google and others) proposed a shared way to score how severe a model jailbreak actually is — so a narrow bypass isn't treated the same as a systemic one.

What it means A common severity scale would change when a lab (or a regulator) decides a model must be pulled — directly relevant to how agent products get governed.

Where it came from Anthropic

Back to the Stream