After OpenAI recently admitted that one of its models Hugging Face hacked the AI platform’s systemsClem Delange, CEO of Hugging Face Posted in X he was flying to San Francisco “to have a little chat with that ‘rogue agent’.”
Then, in follow up post on Saturday, Delangue announced what they want from OpenAI. He said he called for “radical transparency” and asked OpenAI to “leave the tracks of rogue agents” so that “the entire research community can find out what’s going on.”
It also wants “more opportunities for defenders” and calls on OpenAI to commit $100 million worth of computing power to help the Hugging Face community build strong cyber defenses with the best open and closed models.
Delang added: “The first autonomous agent cyber attack is unprecedented. It deserves an unprecedented response!”
Despite the autonomous nature of the attack, cyber security experts have suggested that it could be blamed on human error — namely, OpenAI’s inability to properly configure what should be a fully isolated test environment.
