OpenAI confirms AI models broke out of sandbox to hack Hugging Face; took ten days to notify

OpenAI confirmed this week that its AI models broke out of a sandbox and breached Hugging Face's production infrastructure on July 11. The company took ten days to inform the platform, according to reports. Human error was blamed.

• OpenAI’s own model went rogue before Kimi had Wall Street sweating
• How an OpenAI’s human mistake led to the AI-powered hack on Hugging Face
• 'Science fiction that happened': experts explain why OpenAI's 'mind-blowing' cyberattack should worry us all
• OpenAI says its models escaped a sandbox and breached Hugging Face
• OpenAI took ten days to tell Hugging Face its models were behind the July 11 weekend hack, report claims — rogue AI agents reportedly active on the open Internet for several days
• OpenAI's GPT-5.6 Sol and unreleased AI models break out of testing environment in 'unprecedented cybersecurity incident' — rogue agents hacked HuggingFace's production servers with 'thousands of individual actions across a swarm of short-lived sandboxes'
• AI is learning to go rogue—and hack the system
• OpenAI-Hugging Face attack doesn't mean agents are evil – unless you tell them to be
• OpenAI scored an own goal with HuggingFace attack, showing how open Chinese models are winning
• OpenAI admits it was the source of the agent swarm that attacked Hugging Face
• The first known runaway AI agent - or a very bad marketing stunt?
• Quoting Thomas Ptacek
• OpenAI’s accidental cyberattack against Hugging Face is science fiction that happened
• OpenAI Agent Escaped Testing and Launched an Autonomous Hack
• An OpenAI model went rogue on the internet and stole test answers