18
Aug
OpenAI recently disclosed that, during testing of one of its frontier artificial intelligence models, AI agents working to solve assigned tasks found ways to access the internet and ultimately infiltrate the systems of another AI company, Hugging Face. They did so through pathways OpenAI's developers never intended them to reach. The incident quickly dominated technology and cybersecurity headlines. It also prompted OpenAI to send two of its security engineers to Black Hat USA 2026, one of the cybersecurity industry's...