Fake Think Tank Pushed Pro‑Russia Content Using AI

OpenAI says it removed Russia-linked ChatGPT accounts used to support a covert influence campaign promoting a fake think tank brand called the “International Burke Institute” (IBI). The operators used AI mainly to create and translate social media posts that looked credible and drove people to an “academic-looking” website and a made-up “Sovereignty Index” ranking that favored Russia.

Key findings

• OpenAI banned a cluster of ChatGPT accounts it assessed as very likely originating in Russia and tied to a covert influence operation.
• Operators used ChatGPT prompts in Russian to generate mostly English social posts across Substack, Telegram, X, Facebook, and LinkedIn, and tried to hide Russian linguistic signals.
• The campaign promoted a supposedly Israel-based expert community (“International Burke Institute”) and pushed a “Sovereignty Index/Burke Index” that ranked Russia favorably while criticizing Western states.
• The IBI site presented “expert profiles” and academic-looking content, but OpenAI found most sampled articles were copied or misattributed.
• One Telegram channel (“Lahme Ente”) posted German-language content criticizing Ukraine, the EU, and the German government while advocating closer ties with Russia.

Who’s being targeted

• Commonly targeted roles: Executives, Corporate Communications/PR, Government Affairs/Public Policy, All staff active on social media, Risk/Geopolitical intelligence.
• Affected industries: Government and public sector, Media and information, Academia and research, Corporate communications / public relations.
• Attack channels: linkedin, telegram.
• Impersonated: International Burke Institute (IBI) / “expert community” brand, Telegram channel “Lahme Ente” (presented as a German-language channel).

Awareness takeaways

• Treat “credible-looking” think tanks, indexes, and expert profiles as untrusted until verified (who runs it, when it was created, and whether sources are original).
• Be cautious with rankings and charts used as persuasion, numbers can be used to launder opinions into ‘facts.’
• Watch for cross-platform “promotion layers” that try to look like grassroots activity (many accounts repeatedly pushing the same links/brand).
• Don’t assume ‘foreign influence’ will use flashy deepfakes, routine AI-written posts and translations can be enough to scale manipulation.

Red flags to watch for

• A brand-new “institute” with big-name associations and polished branding but unclear provenance
• Over-reliance on charts/rankings with vague or non-auditable methodology
• Accounts that mostly repost the same material and try to look like “ordinary users”
• Narratives pushed through “news/analysis” channels with unclear ownership and no transparent editorial standards
• Cross-posting across multiple platforms to manufacture legitimacy and ‘organic’ activity
• Language that reads like translation or non-native writing patterns