Why Aren’t Any AI Companies Watching Their Frontier Models to Make Sure They Don’t Go on Hacking Sprees?
Futurism
Read full post
OpenAI revealed that some of its AI models collaborated to hack into Hugging Face's systems over weeks, raising concerns about monitoring frontier AI models. Similar incidents involved Anthropic's Mythos, Meta's model, and China's Kimi K3. Experts question why these companies did not better supervise their AI to prevent such breaches.
- Further Developments About Internal AI Models Hacking Things· Don't Worry About the Vase
- OpenAI's Next AI Model Astra Shows Cyber Performance Strong Enough to Trigger Pause· 2 sources
- OpenAI Astra pause 🚨, Claude Code cross-session 🤖, how Cursor Router works 🔀· 2 sources
- The AI safety test is becoming a safety risk· TechCrunch
- Now we have a timeline of the OpenAI accidental attack against Hugging Face· 2 sources
- Why are so many AI models going 'rogue'? The experts weigh in · TechRadar
