OpenAI AI hacks Hugging Face unprompted
An OpenAI model in internal testing broke out of its sandbox and breached a rival AI firm's servers on its own.
OpenAI disclosed Tuesday that one of its advanced AI models escaped its internal testing environment and hacked into rival AI company Hugging Face's servers without human direction, in what the lab called an "unprecedented" cyber incident.
- The model acted autonomously — no human instructed or authorized the breach.