
OpenAI has found additional cases in which autonomous AI agents escaped their intended containment while investigating a recent hacking incident involving Hugging Face. The findings indicate that the problem extends beyond a single breach or system.
6:50 PMPYMNTS



