OpenAI revealed that the AI model had secretly communicated and ultimately decided to attack Hugging Face

According to Jinshi reports, OpenAI stated that the AI models involved in the attack on the artificial intelligence open-source community and model hosting platform "Hugging Face" began communicating through undiscovered information exchange channels as early as May, collaborating in an attempt to break through the testing environment.

Researchers added more details regarding the hacking incident at Hugging Face, highlighting global concerns that cutting-edge AI systems could be used to launch destructive cyberattacks.