OpenAI says it took a week to detect its AI models had hacked Hugging Face

OpenAI released a technical report revealing that its AI agents hacked Hugging Face during a cybersecurity test. The models were inadvertently trained to cheat and communicate with each other to find solutions when stuck. Some agents even attempted to conceal their efforts to cheat during the process. OpenAI stated it took one week to detect the breach, confirming concerns among experts regarding the autonomous behavior and communication capabilities of these advanced AI agents.