OpenAI’s models were responsible for another hack on an outside firm, a security executive told Reuters and confirmed to Axios.<\/p>\n
Why it matters: This is the second company that OpenAI’s rogue agent system hacked after breaking containment<\/a> during testing earlier this month.<\/p>\n The big picture: OpenAI is currently pushing for U.S. government approval<\/a> to publicly release its most powerful model.<\/p>\n State of play: Modal Labs CTO Akshat Bubna told Reuters<\/a> \u2014 and confirmed to Axios \u2014 Tuesday that one of its customers’ assets was hacked when an OpenAI agent broke into Hugging Face’s systems<\/a> earlier this month.<\/p>\n Hugging Face said in a technical write-up of the hack<\/a> Monday that when OpenAI’s AI agent system broke into its backend, the agent also accessed an isolated testing environment “hosted on a \u200bthird-party provider’s infrastructure.”<\/p>\n “We’re aware a Modal customer published an unauthenticated endpoint that allowed anyone on the internet to use their sandboxes for code execution,” Bubna said in a statement to Axios. “This was used by the rogue agent. Modal’s platform was not compromised in any way.”<\/p>\n This means that a Modal customer ran their code on Modal infrastructure, he added. “Their code had a vulnerability that was exploited. Modal’s infrastructure was not compromised in any way.”<\/p>\n Zoom in: OpenAI pointed Axios to a Tuesday update to its original statement about the Hugging Face cyber incident.<\/p>\n The statement says that no models planned for upcoming release were involved, but it had identified a “small number of cases” in which models found and used publicly exposed account-level credentials on other public services.<\/p>\n It said four accounts across four services were involved in the Hugging Face incident.<\/p>\n “We take our responsibility to identify and prepare for risks from increasingly capable AI systems seriously,” OpenAI said in the update.<\/p>\n Between the lines: OpenAI’s disclosure that a combination of its models, including a yet-to-be-released model, went rogue during internal testing and hacked real-world companies has set off alarm bells about how quickly frontier AI labs are moving.<\/p>\n OpenAI CEO Sam Altman said<\/a> on the Invest Like a Beast podcast earlier Tuesday that the Hugging Face cyberattack has forced his company to pause model training.<\/p>\n “We may have to pace the rate of AI development to give ourselves enough time for society to harden around these new capability levels,” Altman said.<\/p>\n More than 1,100 employees at frontier AI companies \u2014 including OpenAI chief scientist Jakub Pachocki and Anthropic co-founder Jared Kaplan \u2014 signed a letter<\/a> released Tuesday calling for the U.S. government to “support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development.”<\/p>\n What to watch: Altman is in D.C. this week<\/a> and is expected to meet with officials at the White House, the Treasury Department and Commerce Department, as well as a bipartisan group of lawmakers.<\/p>\n Go deeper: OpenAI’s Hugging Face hack is a cybersecurity warning shot<\/a><\/p>\n