New framework grades AI jailbreaks across capability gains, scope, and how easy they are to deploy
Frontier Model Forum releases its Jailbreak Rating Framework to estimate how easy it is for models to escape their cages and how much damage they can do.
A new framework from the Frontier Model Forum (FMF) aims to grade AI jailbreaks, which involve bypassing safeguards. The framework evaluates four key factors: the improvement in capabilities, the scope of the jailbreak, how easy it is to deploy, and how easily it can be discovered. The ratings range from zero to three in each category, with zero indicating the lack of any significant enhancement or posing no threat.
As AI models advance, the framework's grading system adapts, meaning a zero rating in 2026 may carry different implications in 2028.
Written by urgent.news from Tom's Hardware's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
