OpenAI has banned two clusters of ChatGPT accounts it believes were operated from China and used to run covert influence campaigns aimed at American technology and policy debates, according to Tom's Hardware.
The accounts reportedly used AI-generated cartoons to amplify backlash over U.S. data center electricity costs and stir opposition to tariffs. The campaigns appear designed to shape how American audiences think about domestic tech infrastructure — a subject at the heart of ongoing geopolitical rivalry between the U.S. and China.
OpenAI was candid about a certain absurdity in the situation. "It is ironic that the two operations used American AI, rather than Chinese models," the company wrote, according to MSN.
Despite the political sensitivity of the targets, the operations apparently fell short of their goals. As Moneywise reported, OpenAI itself acknowledged the campaigns "didn't do a great job."
The disclosure is part of OpenAI's ongoing effort to surface what it calls influence operations — coordinated attempts to use AI tools to manipulate public opinion covertly, without revealing who is really behind the messaging.
The story matters because it shows that AI tools built in America are now being turned against American public opinion, raising urgent questions about how AI companies can — or should — police geopolitical misuse of their own platforms.