SAN FRANCISCO — OpenAI said it has banned two covert influence operations it links to China that used ChatGPT to generate social-media posts, comments and political cartoons aimed at inflaming American debates over trade tariffs and the build-out of AI data centers.
The disclosures came in the company's June 2026 threat report, described to reporters by Ben Nimmo, principal investigator on OpenAI's intelligence and investigations team.
According to the report and accounts from Reuters and Axios, both operations latched onto debates that were already heated in the United States rather than manufacturing them from scratch. "This was not a case of an influence operation creating a debate," Nimmo told reporters.
One campaign, which OpenAI internally called "Data Center Bandwagon," generated comments and comic strips asserting that AI data centers were driving up electricity prices for ordinary American families, content later posted to X through what OpenAI described as likely inauthentic accounts. A second operation, dubbed "Tech and Tariffs," produced political cartoons criticizing the administration's tariffs and its push for global technology dominance. In one cartoon described by OpenAI, President Donald Trump is depicted swinging a mallet labeled with a tech-dominance slogan into a wall representing the "global future." The same operators, OpenAI said, instructed the model to depict Trump but not
Chinese leader Xi Jinping, and also generated material in Chinese, Italian and Japanese.
OpenAI said the operators relied on virtual private networks to reach ChatGPT from inside China, where the company does not permit access to its models, and that they prompted the system in Chinese while requesting English- and Chinese-language output. The company traced one cluster to people likely working at a private Chinese technology firm doing work for provincial-level government clients, while it said it could not directly attribute the cartoon campaign.
The activity, dating to late 2025 and early 2026, appeared to have little or no real-world effect, OpenAI said. The campaigns nonetheless illustrate how generative AI is becoming a routine tool in digital influence efforts — including ones directed at the AI industry itself.
OpenAI noted the debates the operations targeted are genuinely contested: recent polling has found Americans roughly split on hosting data centers in their communities, and a majority saying tariffs have raised the prices they pay.
Responding to the findings, the Chinese Embassy in Washington said it was not familiar with OpenAI's research but "firmly" opposed what it called groundless attacks against China, adding that Beijing was working to ensure AI serves as a force for good. The episode lands amid an intensifying U.S.-China contest over artificial intelligence that spans chips, data centers and, increasingly, the online debates shaping public opinion about the technology.
© 2026 HNGN, All rights reserved. Do not reproduce without permission.
