The Liberty Daily 
(The Epoch Times)—OpenAI says it exposed and banned two China-linked influence campaigns that used its AI tools to generate content about U.S. technology policy, including debates over data centers and trade tariffs.
Details of the activity were published in the company’s June 2026 Threat Report, released on June 10. The operations involved creating social media posts, images, and political cartoons aimed at American audiences.
The main campaign, which OpenAI labeled “Data Center Bandwagon”, focused on concerns about the impact of AI data centers on electricity prices.
“In this case, the operators attempted to covertly insert themselves into an ongoing American debate about the future of the country’s AI capabilities while hiding who they were and what motivated them,” the report stated.
According to the report, operators generated English-language content claiming that rising power demand from data centers was increasing costs for ordinary American households.
Investigators found that prompts were written in Simplified Chinese and that access to ChatGPT was routed through virtual private networks, as the service is not directly available in China.
The accounts presented themselves as Americans from diverse social and political backgrounds on platforms such as X and Facebook.
The campaign did not invent new grievances but sought to amplify existing concerns. Posts drew on real reporting about electricity markets, power grid capacity, and infrastructure investment before reframing those issues through AI-related narratives.
OpenAI emphasized that the significance of the discovery is not due to the debate itself, which raises legitimate questions on AI policy, but that it represented a covert effort by China to manipulate public opinion while concealing its identity and motivations.
The threat report linked the operation to a social media team at a private Chinese technology company that works with provincial-level government clients. OpenAI did not publish independent evidence establishing direct state control of the activity.
A second network, named “Tech and Tariffs”, produced comments and cartoons criticizing U.S. tariffs and trade restrictions. Materials generated by the operation focused on President Donald Trump, while prompts instructed the AI models not to depict Chinese leader Xi Jinping.
Researchers also connected the campaign to inauthentic X accounts that spread claims alleging ChatGPT user data had been compromised. OpenAI said those allegations were false.
Despite producing large volumes of content, both operations attracted limited authentic engagement. OpenAI’s assessment found no evidence that either campaign meaningfully shifted public opinion.
Ben Nimmo, principal investigator on OpenAI’s Intelligence and Investigations team, said it was “particularly ironic” that foreign actors were using American AI technology to conduct influence operations targeting U.S. audiences.
The findings come as AI infrastructure has become a contentious issue across parts of the United States. Rapid growth in data center construction has raised questions about electricity demand, water consumption, and land use, particularly in states such as Virginia and Texas, where development has accelerated.
The disclosure also fits a broader pattern identified by OpenAI over the past two years. Since the advent of ChatGPT in late 2022, the company has reported disrupting influence operations linked to actors in China, Russia, Iran, and other countries.
Those campaigns have typically involved social media manipulation, deceptive online personas, and attempts to exploit existing political or social divisions.
Previous investigations have repeatedly found that while AI tools can help produce content at scale, many influence operations struggle to gain significant traction with real users.
OpenAI says it releases its threat reports to help governments, researchers, and technology firms identify emerging tactics used in covert online campaigns.
Reuters contributed to this report.
