OpenAI has discovered Chinese-linked accounts using ChatGPT to create fake social media campaigns targeting American opinions about AI data centers. The company published a report detailing how these accounts generated English-language content designed to turn Americans against data center expansion.
The discovery highlights growing concerns about foreign interference in US technology debates. As AI infrastructure becomes central to national competitiveness, foreign actors are apparently trying to influence American public opinion on critical tech policy issues.
OpenAI identified two main groups of suspicious accounts. The first cluster, dubbed the “Data Center Bandwagon” group, focused specifically on data center opposition. These accounts asked ChatGPT to create talking points and images, including comic strips, that highlighted how AI data centers increase electricity demand and drive up consumer bills.
The accounts posed as Americans from various backgrounds on social media platforms. They posted the AI-generated content to make it appear like grassroots opposition to data center development. OpenAI believes these users work for a private Chinese company that serves local government clients.
The group even uploaded files to ChatGPT describing their objectives and strategies for swaying public opinion. The documents included instructions on how to create fake social media accounts without getting detected by platform moderators.
These same accounts also targeted Chinese dissidents abroad. While posing as US-based Chinese immigrants, they asked ChatGPT to generate insults meant to harass political commentators and encouraged online personalities to criticize US policy failures.
The second cluster focused on broader US criticism. These accounts generated comments and images attacking American tariffs and technology policies. They specifically asked ChatGPT to create content portraying the US as unreliable to its allies.
This group requested content in multiple languages – English, Italian, Japanese, and traditional Chinese – to target Taiwanese audiences. They specifically told ChatGPT to keep Chinese President Xi Jinping out of generated images.
The campaigns exploited real concerns about data center impacts. Rising electricity costs near AI facilities are genuine issues affecting American communities. According to Bloomberg, some areas near data centers have seen electricity costs jump 267 percent compared to five years ago due to overwhelming energy demand.
The influence operation’s focus on legitimate grievances makes it particularly concerning. By amplifying real problems, foreign actors can blend authentic concerns with manufactured outrage to shape public debate.
OpenAI admits the campaigns largely failed to gain traction. The company found little authentic engagement with the generated content, suggesting the operation didn’t meaningfully shift public opinion. However, the attempt itself raises important questions about AI tools being weaponized for foreign influence.
The significance lies not in the campaigns’ effectiveness but in their existence. OpenAI notes that 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.”
Interestingly, the Chinese actors chose to use an American AI system rather than domestic alternatives like DeepSeek. OpenAI couldn’t determine the reasoning behind this choice, stating in its report: “We are not in a position to determine what drove this choice.”
The discovery comes as US lawmakers increasingly scrutinize foreign influence in technology debates. Data center expansion has become contentious in many communities, with residents concerned about environmental impacts, energy costs, and infrastructure strain. Foreign actors exploiting these divisions represents a new frontier in information warfare.
