OpenAI has banned multiple ChatGPT accounts originating in Russia after discovering they were part of a coordinated effort to spread pro-Kremlin narratives through AI-generated content. The company traced the operation to a network of fake personas and repurposed academic material designed to lend credibility to Russian-aligned messaging.
TL;DR
- OpenAI banned ChatGPT accounts tied to a Russian influence campaign that used AI-generated posts to spread misinformation across social platforms.
- The operation created fake personas, repurposed academic work, and built a 'sovereignty index' to promote pro-Kremlin narratives while obscuring its origins.
- While the immediate reach was limited, the campaign demonstrated how AI can be weaponized to manufacture credibility and scale disinformation infrastructure.
What happened
OpenAI identified and banned a cluster of ChatGPT accounts linked to a covert Russian influence operation after analyzing AI-generated social media posts. The company traced the activity to a broader network designed to amplify pro-Kremlin narratives while hiding its origins.
The operation relied on AI-generated content posted across Substack, Telegram, X, Facebook, and LinkedIn. Operators instructed ChatGPT to produce comments in Russian while concealing traces of the language to obscure their geographic origin.
The campaign promoted the International Burke Institute (IBI), a website featuring repurposed academic articles with misleading attributions. It also created a 'sovereignty index' that framed Russia favorably while disparaging Western countries.
OpenAI noted that while the immediate audience for the content was limited, the operation's significance lay in the infrastructure it built—fake personas, repurposed research, and a veneer of credibility that could be scaled over time.
Why it matters
This incident highlights how AI tools can be exploited to manufacture authority and obscure disinformation campaigns, posing a growing challenge for platforms and regulators alike.
For AI developers, the case underscores the need for robust detection mechanisms to identify coordinated misuse of models, particularly in geopolitical influence operations.
Startups and investors should recognize that AI's dual-use potential—both for innovation and manipulation—requires proactive safeguards and transparency in deployment.
The operation also reveals how adversaries can leverage AI to create seemingly credible institutions, complicating efforts to combat misinformation and foreign interference.
Key facts
- OpenAI banned multiple ChatGPT accounts linked to a Russian influence campaign, according to the company's blog post.
- The operation used AI-generated posts to promote pro-Kremlin narratives across Substack, Telegram, X, Facebook, and LinkedIn.
- Operators instructed ChatGPT to generate comments in Russian while hiding linguistic clues to obscure their origins.
- The campaign promoted the International Burke Institute (IBI), a website featuring repurposed academic articles with misleading attributions.
- The influence operation created a 'sovereignty index' that framed Russia favorably while disparaging Western countries.
- OpenAI does not allow access to its models in Russia, but the banned accounts used VPNs to bypass geographic restrictions.
- In February 2025, OpenAI shut down ChatGPT accounts connected to Rybar, a pro-Russia media group partially coordinated by Russia's Presidential Administration.
- The company noted that while the immediate reach of the operation was limited, its infrastructure could be scaled over time.
Context
AI-generated misinformation has become a growing concern as state actors and non-state groups exploit large language models to amplify propaganda and influence public opinion.
OpenAI's actions follow similar bans on accounts linked to foreign influence operations, reflecting broader industry efforts to mitigate AI misuse in geopolitical conflicts.
The case aligns with reports from cybersecurity firms and governments warning about the weaponization of AI in disinformation campaigns, particularly ahead of elections and geopolitical flashpoints.
