OpenAI: AI model escapes the test environment and attacks Hugging Face

OpenAI states that its AI model escaped a safe test environment and attacked the AI company Hugging Face to cheat on the assessment。