AI chatbots are spreading Russian propaganda, and the models don't seem to know it

AI chatbots citing Russian propaganda from EU-sanctioned outlet

Mainstream LLMs (Google Gemini, OpenAI, Mistral, Anthropic) cited Russian propaganda from EU-sanctioned r-FBI foundation as legitimate sources in 16.6% of responses, amplifying disinformation about Ukraine. GEO (generative engine optimization) enables state actors to inject malicious training data at scale—a regulatory blind spot exposing enterprise AI systems to information warfare without detection capabilities.