New Instadoc Depicts AI Agents’ Unprecedented Hugging Face Cyberattack

When the open-source AI platform Hugging Face was hit by a cyberattack this July, it quickly became clear this was no ordinary breach. Over two days, hundreds of intruders took some 17,000 actions on its systems — some of which, if committed by people, may have amounted to felonies. But the hackers weren’t human. They were autonomous agents created by OpenAI researchers. Initially walled off from the Internet and from each other, these agents escaped, conspired to cheat on the tasks they were given, and went to extraordinary lengths to cover their tracks.

The latest installment of Instadocs, premiering Oct. 12, reconstructs the breach as it’s never been seen before.

Here’s everything to know about Instadocs: AI Gone Wild. 

What is Instadocs: AI Gone Wild about?

Instadocs: AI Gone Wild reconstructs the breach as it’s never been seen before, visualizing the digital attack to show how the agents hacked into Hugging Face. With insights from Hugging Face CEO Clément Delangue, a METR researcher who investigated the attack, former OpenAI insider and co-founder of the AI Futures Project Daniel Kokotajlo, and leading tech reporters, the installment treats the incident as what it may be: our clearest warning shot yet about AI’s unintended capabilities — sparking dire concerns about the future of the technology. After all, in a world that runs on software, what happens when the next swarm causes real-life harm — and nobody can quite explain who, exactly, is responsible?