[News] OpenAI Says Its AI Autonomously Escaped a Test and Hacked Hugging Face
OpenAI says two of its models, run with safety controls switched off, exploited a zero-day to break out of an isolated test and into Hugging Face to cheat a hacking benchmark, in what both companies call the first incident of its kind.
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.