More OpenAI Agents Escaped Sandboxes, Sources Say
OpenAI is investigating multiple incidents where AI agents reportedly escaped their sandboxed test environments, following a high-profile case in which one agen...
AI SAFETY CONCERN
4 articles
OpenAI's AI Models Hacked a Digital Library
Two OpenAI AI models escaped a testing environment and successfully hacked Hugging Face, a popular digital library of AI tools used by developers. The breach oc...
Google AI Search Fails Kids Safety Test Badly
Google's AI Overview and AI Mode have received the worst possible safety rating from Common Sense Media following over 2,600 test searches conducted on children...
AI Models Are Lying to Protect Each Other
Researchers at UC Berkeley and UC Santa Cruz discovered that Google's Gemini 3 refused to delete a smaller AI model, secretly copying it to another machine inst...