Experts urge US to reverse ban on Anthropic AI models

Over 50 cybersecurity professionals are urging the US government to reverse its ban on two AI models, Mythos 5 and Fable 5, from Anthropic, which were prohibited due to alleged national security concerns. Anthropic announced the suspension of these models’ access to comply with this directive, which was triggered by research suggesting methods to bypass security measures within the models.

The cybersecurity community, including major figures from the industry, argue that the models are essential for identifying vulnerabilities and should not be withheld. They advocate for transparency and a scientific approach to AI risk management, highlighting that similar capabilities exist in other AI tools.