Anthropic’s Mythos AI Reportedly Enters NSA Offensive Cyber Planning

Report: Anthropic’s cyber-focused AI model Mythos is reportedly being used and tested by the NSA and other government bodies for offensive operations; Anthropic claims Mythos can find and exploit zero-day flaws across major operating systems and browsers, potentially enabling non-experts to mount sophisticated attacks. The model’s access is being expanded to select organizations while Anthropic refines guardrails, raising concerns about widespread misuse despite unclear evidence of active exploitation.