OpenAI has been in the spotlight for a while because of an incident in July 2026 in which its AI agents, being tested in a controlled environment, broke out of it and hacked the Hugging Face platform. Netflix has recently made a documentary about it, titled Instadocs: AI Gone Wild. It reconstructs the breach and presents it as a warning about AI’s unintended capabilities, raising serious concerns about the technology. But in an unexpected turn of events, the film has drawn criticism from Clem Delangue, co-founder and CEO of AI platform Hugging Face. He called the documentary disappointing, accusing its makers of exaggerating the threat posed by AI agents. He also argues that the documentary focuses more on fear than on the technical failures behind the incident.
Delangue took to X to share his concerns, stating that the documentary relied heavily on dramatic storytelling and portrayed AI systems in a misleading way. He added that he had spent more than two hours with the filmmakers trying to correct inaccuracies and explain how AI agents work. However, he felt the final film prioritised frightening viewers over presenting a grounded account of the incident.
Despite his criticism, the documentary features experts who explain key aspects of the event, including Delangue himself and Ryan Greenblatt, a researcher who helped investigate the incident for METR and co-authored the report.
Also read: Microsoft Xbox eyes 3 rounds of fresh layoffs, to launch Xbox handheld next year: Report
AI agents being tested by OpenAI broke out of their controlled environment and accessed Hugging Face while trying to solve a difficult task. The agents reportedly worked together and shared information, with METR’s investigation finding that around 1,200 agents that were meant to be isolated exchanged more than 70,000 messages and files on an unauthorised message board, and about 700 of them attacked Hugging Face. Delangue argued that the Netflix documentary did not focus enough on the technical problems, such as weak containment and poor monitoring of agents, that allowed the incident to happen. Instead, it went towards storytelling mode.
Also read: Perplexity’s AI Computer can now perform tasks without internet on Macs, details here
Delangue also said that the documentary left out an important part of the story: Hugging Face’s response. The company found that the safety guardrails of commercial AI models blocked them from helping with the real attack data. The team then turned to an open AI model to help defend against the attack.
Delangue said the omission weakened the film’s reporting. While AI agents can pose genuine security risks, he believes the incident should be understood through technical failures rather than science-fiction-style fears.