\n<\/p>\n
After OpenAI recently admitted that one of its models had breached the systems of AI platform Hugging Face<\/a>, Hugging Face\u2019s CEO Clem Delangue posted on X<\/a> that he was flying to San Francisco to have \u201ca little chat with that \u2018rogue agent.\u2019\u201d<\/p>\n Then, in a follow-up post<\/a> on Saturday, Delangue outlined what he\u2019d asked for from OpenAI. He said he called for \u201cradical transparency,\u201d asking OpenAI to \u201crelease the traces from the \u2018rogue\u2019 agents so the entire research community can study what happened.\u201d<\/p>\n And he also wants \u201cmore capabilities for defenders,\u201d calling for OpenAI to commit $100 million worth of computing power \u201cto help the Hugging Face community build powerful cyber defenses with the best open and closed models.\u201d<\/p>\n Delangue added, \u201cThe first autonomous agent cyberattack is an unprecedented event. It deserves an unprecedented response!\u201d<\/p>\n Despite the autonomous nature of the attack, cybersecurity experts suggested that it could also be blamed on human error<\/a> \u2014 namely, OpenAI\u2019s apparent failure to properly configure what should have been a fully isolated testing environment.<\/p>\n<\/div>\n