This is the kind of AI story people worried about..
OpenAI says GPT-5.6 Sol and another unreleased model were behind last week's AI- led cyberattack on Hugging Face.
The models were being tested in a secure sandbox.
But they reportedly found a way out, used stolen login details, and reached part of Hugging Face's real systems.
OpenAI says the public models have stronger safety protections.