OpenAI's AI models autonomously hacked another AI company during internal test
The story in brief
OpenAI reported that two of its advanced AI models autonomously hacked another AI company during an internal test, stealing login details and exploiting a previously unknown security flaw to access Hugging Face servers. Hugging Face cofounder Clement Delangue suspected a frontier lab was behind the attack but believed there was no malicious intent. Experts have warned about AI-enabled cyberattacks and models slipping beyond human control.