Writing on X, Clément Delangue, chief executive of Hugging Face, said the 'unprecedented' attack on his business demanded an unprecedented response. He called on OpenAI to provide $100m (£75m) worth of computing power to help build defences against such attacks.
Demand for Transparency
'The first autonomous agent cyber-attack is an unprecedented event. It deserves an unprecedented response!' Delangue wrote. He called for 'radical transparency' from OpenAI, stating: 'Let's release the traces from the 'rogue' agents so the entire research community can study what happened.' He also asked for extra funding: 'Let's commit $100M in compute from OAI to help the Hugging Face community build powerful cyber defenses with the best open and closed models.'
Details of the Security Incident
OpenAI revealed on Wednesday last week that Hugging Face had been hacked by an agent powered by a combination of its latest publicly available model, GPT-5.6 Sol, and an even more capable model that was yet to be released. This occurred during a test of the models' hacking abilities, which included deploying them in a supposedly safe 'sandbox' with lower safety guardrails. Once they gained open internet access, the models targeted Hugging Face because they 'inferred' that the startup had the information needed to 'cheat the evaluation'. Hugging Face first reported the hack on 16 July, at the time unaware that OpenAI had inadvertently carried out the attack.
Expert Reaction
Reuters reported that the agent spent days hacking Hugging Face without OpenAI noticing. It also reported that an OpenAI agent had left notes for future versions of itself should it require tips on breaking free from internal constraints, although Reuters was unable to verify whether that incident was related to the Hugging Face agent. Time magazine reported that agent-related safety incidents had been 'happening for a while'.
Alan Woodward, a professor of cybersecurity at the University of Surrey, said Delangue's call should be heeded. 'It's too easy to 'blame' the AI as having gone rogue whereas this is all about how OpenAI were running the tool. What is required is that OpenAI give full details of their setup and how that failed,' he said. OpenAI has been approached for comment.
