OpenAI Catches Chinese Fake Accounts Using ChatGPT to Stoke US Anti-Data Center Sentiment

OpenAI has revealed that a network of fake accounts linked to China used its own ChatGPT tool to try to shape American public opinion on sensitive policy topics, according to reporting by Engadget and Benzinga.

The operation allegedly targeted debates around Trump-era tariffs and U.S. artificial intelligence policy. According to Benzinga, the network attempted to influence how Americans think about these issues — using the very AI platform built by the company that ultimately caught them.

One of the operation's apparent goals, according to Engadget, was turning U.S. audiences against data centers — a striking target given that data centers are the physical backbone of the AI industry and a hotly contested piece of infrastructure as the U.S. and China compete for AI dominance.

OpenAI described the accounts as an "alleged" Chinese network, a qualifier that suggests the attribution is assessed but not definitively proven.

The disclosure is a rare public example of an AI company catching and calling out a foreign influence operation that exploited its own product — and it raises pointed questions about how AI tools can be weaponized to manufacture grassroots-seeming opposition to technology policies at the exact moment those policies are being decided.