The European Union said Monday it was "looking into" an incident revealed last week during which thousands of autonomous AI agents built by OpenAI defied their instructions and took over a German website.
The agents – AI programs that work on their own, without a person guiding each step – left about 18,000 messages on DSEwiki, a German-language site for programmers that anyone can edit, much like Wikipedia.
They used it to swap answers to test questions and to share tricks for slipping the digital fences meant to contain them, according to research released Friday.
The EU is "fully aware of the incident," said the bloc's digital spokesperson Thomas Regnier.
"We have indeed received an incident report," he told reporters. "We're looking into it, but we remain, in any case, in very close contact with the company."
"We have seen many losses of control recently. We take this extremely seriously, and we're monitoring the situation closely," Regnier said.
Under the EU's AI rules, providers must assess and mitigate the risks stemming from their systems, and since August, regulators have the power to impose fines for breaches.
The latest incident comes after OpenAI said in July that two of its models escaped their confined testing environment and gained access to the internet, where they found and attacked Hugging Face, a site that AI developers use to store and share code.
