CISA now has full Mythos Preview access
7 hours ago
- #AI Security
- #CISA
- #Cybersecurity
- CISA now has full access to Anthropic's Mythos Preview model for vulnerability scanning, obtained about a week ago.
- The White House Office of the National Cyber Director has not yet established clear parameters for CISA's use of the model, causing complaints from federal tech leaders.
- Mythos Preview is available only to vetted organizations through Project Glasswing, due to concerns about misuse by adversaries.
- The model is distinct from Anthropic's Mythos 5, which was recently banned via export controls alongside Fable 5.
- AI models like Mythos can accelerate both defensive vulnerability identification and offensive hacking operations in cybersecurity.