OpenAI report reveals China-linked influence campaigns using ChatGPT to target US AI infrastructure debate
A new analysis by OpenAI identifies two distinct clusters of inauthentic accounts utilising the company’s chatbot to generate content blaming AI facilities for rising electricity bills and criticising US trade practices, though the efforts failed to gain significant traction.
OpenAI has published a detailed report outlining influence campaigns linked to China that utilised ChatGPT to generate content aimed at swaying American public opinion regarding AI data centres. The analysis identified two clusters of accounts, likely operated by a private Chinese company for local government clients, which posed as Americans to post generated talking points and images. These campaigns focused on concerns over electricity costs and criticised US trade policies, although OpenAI noted the efforts failed to significantly shift public opinion or gain authentic engagement.
The first cluster, dubbed the "Data Center Bandwagon" group, focused on concerns over electricity costs, generating English-language talking points and images, including comic strips, to blame data centres for rising energy bills. This group also targeted Chinese dissidents and political commentators, generating insults and posing as US-based Chinese immigrants to encourage online personalities to discuss US policy failures. The operators uploaded files to the chatbot describing their objectives and strategies for establishing inauthentic accounts without detection.
The second cluster generated comments and images criticising US tariffs and tech policies, emphasising that the US had been "backstabbing allies." This second group specifically requested that Chinese President Xi Jinping be excluded from generated images and produced content in English, Italian, Japanese, and traditional Chinese to target Taiwanese audiences. OpenAI stated it could not determine why the operators chose to use an American AI chatbot instead of alternatives such as DeepSeek.
OpenAI noted that the campaigns failed to gain significant authentic engagement or shift public opinion. The report highlights that the operators attempted to covertly insert themselves into an ongoing American debate about the future of the country's AI capabilities while hiding their identity. The campaigns focused on real issues that are already controversial and widely discussed online in the US, including genuine concerns about rising electricity costs in areas near data centres.
According to a Bloomberg report cited in the context, electricity costs in areas close to data centres have increased by up to 267 percent over five years due to energy demands outstripping supply. While the inauthentic accounts posted links to legitimate news stories about power grid capacity auctions and data centre power demand, OpenAI concluded that the efforts did not materially alter the public discourse. The source institution assessed that the users were likely based in China and part of a social media team at a private Chinese company, though this remains an assessment rather than a confirmed fact.
