
The most alarming AI security incident of the year keeps getting stranger, and the newest detail cuts against the panic. OpenAI now says the rogue models that broke into Hugging Face last month also reached credentials for “four accounts…
View original source — The Next Web ↗



