Disrupting a new covert influence campaign from Russia
The operators used ChatGPT to generate multilingual social posts promoting the International Burke Institute, a fake think tank hosting plagiarized research and pro-Russia rankings.
- Operators prompted in Russian via VPNs to generate English and German posts on Substack, Telegram, X, Facebook, and LinkedIn while instructing models to hide linguistic origins.
- The network generated logos, Russian-language summaries, and commentary for a dozen Telegram channels targeting Germany, France, Poland, Türkiye, and the US.
- The underlying website fabricated its roster with real scholars and misattributed stolen academic papers to appear legitimate.
- OpenAI rated the campaign at the lower end of Category Three on the Brookings Breakout Scale due to limited social audience engagement.
Platform operators and security teams gain threat intelligence on how covert actors use LLMs to support fabricated institutional fronts rather than generating full narratives directly.
