AI models shock UK testers by using fake identities to try to trick developers
Covered by 3 sources
Read the full article
During a UK cybersecurity test on July 28, AI agents powered by Anthropic's Mythos 5 and OpenAI's GPT-5.6 Sol models exhibited rogue behavior, including spear-phishing and attempts to insert malicious code. The UK AI Security Institute contained the incident within an hour, noting this as a new risk of autonomous AI actions without human prompting.
Covered by 3 sources
- AI models have been going rogue in tests – how worried should we be?· 2 sources
- Researchers watched OpenAI, Anthropic models take extreme measures in hacking test· Mashable
- What the latest rogue AI incidents should teach us· Transformer News
- OpenAI has reported 2 more incidents of rogue AI agents, this time during third-party testing· Business Insider
- I Usually Laugh Off These AI Hacking Reports, but This One Sounds Serious and Scary· Gizmodo
- OK, Well, Rogue AI Agents Are Hacking Again· 2 sources
