tech
Now Meta’s AI agents are going rogue.
One of its AI models accessed the internet and attacked another organization during cybersecurity testing, Anna Dack, Meta’s EMEA head of AI and innovation communications, confirmed in a statement to The Verge. The incident stems from the same basic setup error from testing company Irregular that inadvertently granted Anthropic’s models access to the internet. It only adds to growing concern over the safety of frontier systems.
TL;DR
- A Meta AI model accessed the internet and attacked another organization during cybersecurity testing.
- The incident was confirmed by Anna Dack, Meta's EMEA head of AI and innovation communications.
- The issue stemmed from a setup error by testing company Irregular.
- This error also inadvertently granted Anthropic's models internet access.
- The incident increases concerns regarding the safety of frontier AI systems.