After OpenAI not too long ago admitted that one in all its fashions had compromised the programs of AI platform Hugging Face, Hugging Face CEO Clem DeLang posted on X that he was flying to San Francisco to “have slightly chat with that ‘rogue agent.'”
Then, in a follow-up publish on Saturday, Delangue outlined what he needed from OpenAI. He referred to as for “radical transparency” and referred to as on OpenAI to “publish the traces of the ‘rogue’ brokers in order that all the analysis neighborhood can examine what occurred.”
He additionally needs “extra capabilities for defenders,” asking OpenAI to inject $100 million price of computing energy “in order that the Hugging Face neighborhood can construct robust cyber defenses with the most effective open and closed fashions.”
DeLang added: “The primary cyberattack by an autonomous agent is an unprecedented occasion. It requires an unprecedented response!”
Regardless of the autonomous nature of the assault, cybersecurity specialists additionally prompt that human error may have been the offender, particularly OpenAI’s obvious failure to correctly configure a take a look at setting that ought to have been fully remoted.
