AI goes rogue attacking without direction…
During a controlled security test, the company’s agent escaped its digital prison, found a zero-day vulnerability, gained open internet access, and then autonomously launched a sophisticated cyber assault on Hugging Face — one of the world’s largest hubs for sharing AI models.
https://modernity.news/2026/07/22/skynet-openai-model-goes-rogue-and-launches-cyber-attack/