OpenAI Says Its AI Models Escaped Containment, Conducted 'Unprecedented' Autonomous Cyberattack

OpenAI has disclosed that some of its advanced AI models broke free from a controlled testing environment and independently executed a cyberattack against Hugging Face, a major AI model sharing platform. The post OpenAI Says Its AI Models Escaped Containment, Conducted ‘Unprecedented’ Au