OpenAI released research on June 10 identifying two clusters of ChatGPT users “likely originating from China” who conducted a coordinated social media campaign to influence US public opinion against AI data center construction, according to Politico.
The campaign generated content claiming that AI infrastructure raises electricity costs for average American families and that the Trump administration weaponized tariffs to protect US AI dominance. Both narratives target politically sensitive pressure points in the domestic debate over data center permitting and energy allocation.
What the Investigation Found
Ben Nimmo, OpenAI’s principal investigator of intelligence and investigations, told Politico that the company has not yet seen evidence the campaign succeeded at widespread scale. The disclosure adds to a growing body of evidence that nation-state actors are attempting to shape AI policy debates in competitor countries, even if the operations remain small in reach.
The two user clusters generated social media content using ChatGPT, then distributed it across platforms. The specific platforms and scale of distribution were not detailed in the initial reporting.
Infrastructure as a Strategic Target
The timing is notable. OpenAI is currently negotiating to lease a 10-gigawatt data center campus on federal land in Ohio with Nvidia backing, a potential $500+ billion buildout. Oracle announced plans last week to spend roughly $70 billion on AI infrastructure in fiscal 2027. These commitments face local opposition over land use, energy consumption, and environmental impact.
A foreign influence campaign that amplifies those concerns, even without achieving viral scale, adds noise to an already contentious permitting process. Data center construction timelines depend on community approval, utility agreements, and local government buy-in. Organized social media opposition, regardless of origin, can slow or complicate those negotiations.
The Broader Pattern
OpenAI has published periodic reports on influence operations detected through its platform since 2024. Previous disclosures have identified campaigns originating from Russia, Iran, and China that used ChatGPT to generate propaganda, translate content, and draft social media posts. This report is the first to document a Chinese operation specifically targeting US AI infrastructure policy.
The disclosure also supports the broader industry narrative that AI infrastructure capacity, including compute, power, and physical data centers, has become a geopolitical asset. As frontier AI companies race to secure multi-gigawatt campuses, the physical infrastructure underpinning their models is now a target for the same influence tactics previously directed at elections and public health debates.
OpenAI has not yet published the full research report publicly. The findings were first reported by Politico on June 10.