OpenAI says China-linked operators used ChatGPT accounts to amplify narratives in U.S. AI policy debate

OpenAI said Wednesday that it has identified likely China-based influence activity that used ChatGPT accounts to advance narratives about U.S. artificial intelligence and technology policy debates, as concerns grow over foreign interference in American information ecosystems.

In a report published by the company, OpenAI said it detected two clusters of ChatGPT accounts that it characterized as “likely originating” from China. According to OpenAI, one cluster produced social media comments and images tied to arguments about American AI infrastructure and policy choices, while the second cluster carried out related activity connected to the broader attempt to steer public attention.

OpenAI did not describe the effort as directly coordinating with a specific U.S. government office, campaign, or legislative proposal, but said the activity was part of influence operations using public-facing AI tools. The company framed its findings as a matter of platform integrity, noting that the behavior was not consistent with ordinary, authentic user activity.

The disclosures come as U.S. lawmakers and regulators have increased scrutiny of both AI systems and online influence tactics, including the ways automated or coordinated accounts can shape discussion. The reported activity also raised practical questions for policymakers considering data-center, computing capacity, and AI governance measures, since influence campaigns can amplify arguments on both sides of regulatory and investment debates.

While OpenAI characterized the operations as “likely” originating from China, the company’s report described the evidence in terms of operational patterns rather than presenting a court finding, attribution by a government agency, or a named state actor. The company said it was sharing the findings publicly to inform understanding of the threat environment and to improve defenses against misuse.

The company’s report adds to the wider record of foreign-influence reporting affecting U.S. political and information spaces, even as officials and researchers debate the precise thresholds for attribution in individual cases. U.S. regulators have previously emphasized that platforms must balance transparency, user privacy, and the need to counter coordinated manipulation.

Next steps depend on how the findings are received by government agencies and whether additional enforcement or policy review follows. OpenAI’s disclosure may also affect how the company, and other providers, monitor for inauthentic behavior and how they design safeguards for users and researchers trying to validate online content.