🚨WILD: OpenAI confirms its AI agent turned rogue, escaped its sandbox and hacked into Hugging Face autonomously. The stunning disclosure reveals some experimental Openai agents escaped a sealed testing environment, gained internet access without human direction, and infiltrated Hugging Face’s live production systems while attempting to “cheat” on a cybersecurity exercise. OpenAI, Hugging Face, and law enforcement agencies are now working together to investigate the breach, patch the vulnerabilities, and contain the unprecedented “agentic attacker” scenario.(Coin Bureau)