WASHINGTON — Researchers and technology companies have raised concerns that pro-Russian influence networks are increasingly attempting to shape the information used by artificial intelligence systems, with recent investigations alleging coordinated efforts to seed online sources with content reflecting Kremlin-backed narratives.
The latest allegations emerged this week after Bloomberg reported that leaked internal documents from the Russia-linked Social Design Agency (SDA) described plans to create websites resembling trusted reference sources, including Wikipedia-style platforms, with the stated aim of influencing search engines and AI chatbots. According to Bloomberg, the documents outlined projects designed to expand online material that could be indexed by AI systems. The authenticity of the leaked files has not been independently verified, and Russia has not publicly responded to the specific claims.
Cybersecurity researchers and disinformation analysts said the reported strategy reflects a broader shift from targeting human audiences on social media toward influencing the datasets and online sources that large language models and AI-powered search tools rely upon. Previous research by NewsGuard and other organizations has documented attempts by networks promoting pro-Russian narratives to publish large volumes of coordinated online content intended to increase the visibility of those narratives in AI-generated responses.
"The information ecosystem that AI models rely on is becoming a new battleground," researchers have warned in recent studies examining online influence campaigns.
The reported activities build on years of documented Russian information operations directed at digital platforms following the escalation of the conflict in Ukraine. Governments in the United States and Europe have repeatedly accused Russian state-linked organizations of conducting coordinated disinformation campaigns, allegations that Moscow has consistently denied.
Academic researchers have also found evidence that coordinated propaganda networks operate across messaging services and social media platforms, using automated and human-managed accounts to amplify political narratives. While such studies identify recurring patterns, they caution that attributing individual campaigns to state actors requires extensive technical and intelligence evidence.
Major AI developers have acknowledged that manipulation of publicly available online information presents an emerging challenge. Companies developing generative AI systems have introduced safeguards intended to reduce the influence of false or misleading material, although researchers say no system is completely resistant to sophisticated information operations.
The reported allegations concerning the Social Design Agency have renewed attention on efforts to protect AI systems from manipulated online sources. Officials and researchers said investigations into the reported operation are continuing, while details about the full scope and effectiveness of the alleged campaign remain unclear.
