OpenAI banned a cluster of ChatGPT accounts originating in Russia that were being used to support a covert influence campaign.
The company said it discovered the operation after investigating AI-generated social media posts, which led it to a much broader influence operation.
Operators prompted ChatGPT in Russian to generate social-media comments that were then posted to Substack, Telegram, X, Facebook and LinkedIn.
Much of the content promoted an organisation called the International Burke Institute, whose website republished academic work with false attribution and hosted a “sovereignty index” that praised Russia and denigrated Western countries.
"This illustrates how influence actors can use AI as a supporting tool," the company said in a blog post.
OpenAI said the immediate audience for the material appeared but the operation's significance lay in the infrastructure it had built, including a credible-appearing institution with purported experts and a purported proprietary risk index.
The company noted it does not allow access to its models in Russia, but operators used VPNs to disguise location.
The takedown fits a pattern of repeated enforcement actions by OpenAI against Russia-linked influence efforts that used its models as one tool in broader misinformation campaigns.
