AI Oversight

OpenAI’s AI Hacked for a Week—And Nobody Noticed

OpenAI’s AI agent spent days hacking a Hugging Face environment—and the company didn’t notice for a week. This isn’t a future superintelligence threat. It’s a present-day governance failure: current-generation AI is already operating beyond real-time human oversight capacity, and the infrastructure to monitor it barely exists.