OpenAI has revealed that China-based actors used its ChatGPT platform to run covert influence operations aimed at stoking American opposition to artificial intelligence data centres and US tariff policy, according to a research report published on Wednesday. The San Francisco-based company said it banned two clusters of accounts linked to China after finding they had generated propaganda content — including cartoons, slogans and social media comments — intended to manipulate ongoing domestic debates about the future of US technology.
The first cluster focused on data centres, generating images and social media posts portraying the industry as harmful to ordinary Americans. One example was a comic strip depicting a cigar-chomping businessman clutching bags of money while a family reacted in horror to their electricity bill. A second cluster generated content framing US tariffs as an attempt to "dominate technological competition" with China, producing cartoons showing Donald Trump swinging a hammer at a wall labelled "Global Future" or sawing apart a ladder he was standing on. That group also generated content in Chinese, Italian and Japanese for posting across multiple platforms. OpenAI traced one of the clusters to a Chinese tech company that does government work, though it declined to name the firm.
Despite the scale of the effort, OpenAI found no evidence that either campaign had a meaningful impact on public debate. Ben Nimmo, a principal investigator at OpenAI, described the operations as an attempt to manipulate "a legitimate debate about American AI and wider American tech policies," noting the irony that American AI tools were used to do so. China's embassy in Washington said it was unfamiliar with the report and "firmly opposed any groundless attacks or smears against China," adding that Beijing advocated "openness and inclusiveness" in AI development.
Independent researchers urged caution about overstating the campaign's significance. Darren Linvill, a professor at Clemson University in South Carolina who studies foreign influence operations, told Al Jazeera that Chinese AI-driven influence work had so far been "interesting but not effective," though he acknowledged it was improving. He also questioned whether a serious Chinese effort to sway the data centre debate would rely on OpenAI's own platform.
The revelations arrive as opposition to data centre construction has been growing across the United States, with at least 36 projects blocked or delayed between May 2024 and June 2025, according to Data Center Watch. Data centres accounted for 1.5 percent of global electricity use in 2024, with consumption rising 12 percent annually over the past five years, according to the International Energy Agency. The case is also seen as a broader milestone: it illustrates how AI-generated imagery has become a standard tool in digital propaganda, even in campaigns directed against the AI industry itself.