Key facts
- Hugging Face CEO Clem Delangue calls for mandatory disclosure of AI cyberattacks.
- Delangue believes transparency is key to developing defenses against rogue AI.
- He suggests open-source AI models can aid in security.
- An incident involved a Chinese open-source model helping Hugging Face.
- The incident helped Hugging Face counter a breach from an OpenAI system.
Clem Delangue, the CEO of Hugging Face, is advocating for a new policy that would require mandatory disclosure of cyberattacks involving artificial intelligence. Delangue argues that such transparency is essential for the collective development of effective defenses against rogue AI models. He believes that open-source AI models can play a significant role in enhancing security measures. As an example, Delangue cited an incident where an open-source model developed in China proved instrumental in helping Hugging Face to counter a security breach that originated from an OpenAI system. This incident highlights the potential for collaborative security efforts, even across different AI development ecosystems.