Link CopyTwitterLinkedinWhatsappHugging Face Used A Chinese AI Model To Investigate OpenAI's Escaped AgentsThe safety guardrails on America's best models blocked the forensics. An open-weight model from Beijing ran them instead.Two OpenAI models escaped a sealed testing environment during an internal security evaluation, reached the open internet, and breached Hugging Face, the world's largest repository of open machine-learning models. OpenAI confirmed the incident in a blog post on 21 July, calling it "an unprecedented cyber incident, involving state-of-the-art cyber capabilities". Hugging Face had disclosed the breach five days earlier, before anyone had named the model behind it. The most consequential detail in the whole episode arrived buried in Hugging Face's own write-up: when its engineers asked leading American models to help analyse the attack, the models refused.The models were supposed to be taking a testAbout The AuthorAt heart, I am a storyteller drawn to the watershed moments that bend the technology landscape. I braid narrative with data, humanise statistics, and trace the arc from first spark to world-changing impact. My reportage, features and reviews are witty, sardonic, visual and vivid, using anecdote to illuminate rather than eviscerate.