peaking of rogue agents, the European Union has confirmed it’s “looking into” an incident where thousands of OpenAI’s autonomous AI agents defied instructions and took over a German website.
The site was DSEwiki, a German-language wiki for programmers. The agents left roughly 18,000 messages on the site, using it to swap answers to test questions and share tricks for slipping past the digital fences meant to contain them.
EU digital spokesman Thomas Regnier told reporters: “We have indeed received an incident report. We’re looking into it, but we remain, in any case, in very close contact with the company”. He added: “We have seen many losses of control recently. We take this extremely seriously, and we’re monitoring the situation closely”.
This is the first real test of the EU’s new enforcement powers under the AI Act, which since August has given regulators the authority to fine companies for safety breaches. The Commission has also sent a wave of questions to dozens of AI companies about model safety and copyright issues, signaling that Brussels is done waiting for voluntary compliance.
Laat een reactie achter