Anthropic’s AI Security Claims Tested After Mythos Model Breach
A breach reveals that Anthropic’s restricted AI model Mythos, marketed as too dangerous for public release, was accessed by unauthorized users shortly after its controlled rollout began.