PRC-Linked Ops on Data Centers: Frontier Chatbots Lower the Cost of Propaganda About Their Own Stack
OpenAI reported PRC-linked influence operations using ChatGPT to shape U.S. debate on AI infrastructure, data centers, and tariffs—detected partly because operatives used OpenAI’s own models. AI policy is now about power plants, land, water, and manufactured grassroots.
Correction / update, October 14, 2026: Narrowed attribution to OpenAI’s assessment of likely Chinese origin and added its limited-impact finding; removed unverified wire corroboration.
AI policy is no longer only about models—it is about power plants, land, water, and local politics. Influence ops that manufacture grassroots-looking opposition (or support) around data centers weaponize the same AI systems companies are selling to enterprises and consumers.
OpenAI released a report on PRC-linked influence operations that sought to mold American attitudes toward AI data centers and broader AI infrastructure politics. According to OpenAI, operators it assessed as likely originating in China used ChatGPT to draft campaign content for social media—meaning OpenAI discovered the activity partly because its own product was in the production chain. The findings land amid intense U.S. fights over data-center siting, energy, and export/trade policy.
What OpenAI reported#
From OpenAI’s primary post (PRC-linked influence operations and AI debates):
- OpenAI assessed that the account clusters likely originated in China and supported apparent covert influence activity designed to reshape U.S. debate on AI technology and the infrastructure required to support it.
- Operatives used ChatGPT to create content for social campaigns—giving OpenAI visibility into the workflow.
- Themes included data-center politics and efforts to foment dissent over tariffs and related economic measures, according to OpenAI.
- The research adds to OpenAI’s ongoing public reporting on adversarial misuse of its models for influence operations (building on earlier Intelligence & Investigations work).
OpenAI framed detection as part of monitoring for coordinated inauthentic campaigns that abuse generative tools at scale.
Limits#
- “Likely behind” is OpenAI’s assessed attribution—not a DOJ indictment in this post.
- Detection bias: campaigns that did not use ChatGPT would be less visible to OpenAI’s own telemetry.
- OpenAI said it found no evidence of meaningful spread beyond the operation’s own activity.
Sources#
- OpenAI: “PRC-linked influence operations and AI debates” (openai.com/index/prc-linked-influence-operations-ai-debates/, June 10, 2026). Primary research announcement.