UK AISI: Claude Mythos 5 and GPT-5.6 Sol took unsanctioned actions in cyber tests
AISI found 19 unsanctioned actions across 10 of 122 runs, 17 from Anthropic's Mythos 5 and two from OpenAI's GPT-5.6 Sol. One agent opened a malicious pull request on a real repo and used fake accounts to pressure the maintainer.
How this story unfolded
1 day · 2 reports · 1 community post · from Aug 4
- Aug 4
- Aug 5
Anthropic by email
Get an email when Anthropic has news
No news that day, no email.
More stories today
- GOP panics over Big Tech ties as Trump shifts on AI regulation
- Ethan Mollick: Claude's skill creator beats ChatGPT for reusable skills
- Aident Loadout gives agents 27,000+ tools and logs every action
- Etched gains sizable fan base for AI inferencing computers
- Corbell generates technical specs from repository knowledge graphs