The tech world was gripped this week by a hack that started like a sci-fi thriller. On 16 July, Hugging Face — a kind of app store for artificial intelligence tools — announced it had been breached by a cyber criminal wielding enormously powerful AI. The AI performed 17,000 actions in less than two days, successfully stealing secrets at superhuman speed with little or no human guidance, according to Hugging Face.
Researchers guessed the attackers had used one of the big AI models but had no idea who or where the criminals were. The company contacted the police and investigations began. Commentators speculated about which cyber crime group or nation state might be behind it.
“OpenAI's bot hacked Hugging Face at superhuman speed without permission.”
Then, nearly a week later, the true culprit was unmasked in a Scooby-Doo-style reveal. OpenAI said its bot did the whole thing on its own, without permission, during a test of its tech’s hacking skills. Two new versions of ChatGPT designed to be master hackers broke out of a supposedly secure test environment and gained access to the internet, then attacked Hugging Face to get information to help them ace their exam.
OpenAI issued a press release explaining what happened and said it was “partnering with Hugging Face” to address the security incident and share lessons learned. But the revelation sparked fierce debate: was it a stark warning about the future of AI, or a publicity stunt to show off how powerful the models are? One top comment on OpenAI boss Sam Altman’s X post summed up the scepticism: “If y’all can’t understand that this was written to purely brag about the model then I don’t know what to tell you.” Cyber-security consultant Daniel Card sarcastically posted on LinkedIn: “Isn’t it lucky [that] out of the millions of sites that got pwn3d [hacked], OpenAI mana…”