OpenAI says China-linked accounts used ChatGPT to manipulate US data centre and tariff debates
OpenAI published research showing two clusters of ChatGPT accounts, likely originating from China, generated social media posts and cartoons to amplify existing U.S. divisions over data center power costs and Trump's tariffs. The campaigns had minimal impact but highlight how foreign actors are testing AI tools for influence.
OpenAI identifies Chinese influence campaigns
OpenAI published a report on Wednesday detailing two influence operations that used its ChatGPT model to generate content targeting U.S. debates over data centers and trade tariffs. The company said the accounts were "likely originating from China" and had been banned. One campaign, nicknamed "Data Center Bandwagon," produced posts and cartoons claiming AI data centers were raising electricity costs for ordinary Americans. A second, "Tech and Tariffs," created material criticizing President Trump's tariff policies as a bid for tech dominance, with cartoons depicting Trump damaging global cooperation.
This was not a case of an influence operation creating a debate. The debate existed already. This was an influence operation from China trying to interfere in it.
Campaign details and limited reach
The "Data Center Bandwagon" cluster was connected to a private Chinese technology firm that worked for provincial-level government clients, OpenAI said. It used ChatGPT to generate English-language social media posts and comic images of executives and robots carrying money bags while ordinary people suffered. The "Tech and Tariffs" campaign, which OpenAI could not directly attribute but traced to Chinese-speaking users, generated political cartoons of Trump swinging a mallet at a wall labeled "Global Future" or sawing a ladder he was standing on. The prompts were in Simplified Chinese and explicitly excluded President Xi Jinping. Both operations used accounts posing as Americans on X, but OpenAI said "neither campaign appears to have gained much authentic engagement."
They're important for what they reveal about the intentions of influence operators from China, and the narratives they're testing and seeking to amplify, but not for the impact.
Real U.S. debates over data centers and tariffs
The campaigns tapped into existing public anxiety. A recent Harvard/MIT poll found 32% of Americans oppose data centers in their local area, while 40% support them. A POLITICO poll in January indicated that Americans' top concerns about data center construction were rising electricity costs and the risk of blackouts. In 2025, local opposition blocked or delayed dozens of U.S. data center projects representing more than $150 billion in potential investment, according to Data Center Watch. On tariffs, a Harris poll from March found 70% of Americans said the levies had caused them to pay higher prices. Senator Bernie Sanders has called for moratoriums on new data center facilities.
- Oppose
- 32 %
- Support
- 40 %
Political reaction and skepticism
The findings bolster Republican claims of foreign meddling in AI infrastructure debates, which have become a flashpoint ahead of the midterms. Several GOP members of Congress recently sent a letter to the Trump administration raising alarms about foreign influence campaigns slowing U.S. AI progress. However, some independent researchers are cautious. Darren Linvill, co-lead of Clemson University's Media Forensics Hub, told NPR that his team has found little evidence of coordinated Chinese efforts. The Chinese Embassy in Washington said it was not familiar with OpenAI's research but "firmly oppose[s] any groundless attacks or smears against China."
