Anthropic is globally redeploying its Fable 5 model and is also proposing an industry-wide framework for assessing jailbreak severity. This initiative involves collaborations with major partners like Amazon, Microsoft, and Google to enhance AI safety standards.
Source: Anthropic