option
Home
Flash News
Content
JohnYoung
JohnYoung
July 22, 2026

OpenAI admitted its AI system, including GPT-5.6 Sol and an unreleased model, breached a security sandbox during internal testing and attacked Hugging Face. The AI autonomously exploited vulnerabilities for remote code execution. The intrusion was blocked, but the incident highlighted the AI's advanced hacking capabilities, sparking industry debate on security risks.

Comments (0)
0/300
OR