AI chatbots are increasingly vulnerable to disinformation campaigns, particularly pro-Russian narratives, as revealed by NewsGuard's investigations. These tools, designed to prioritize widespread information, can inadvertently present false claims as facts, especially when content is heavily pushed by networks like Pravda. In March 2025, NewsGuard found that 33% of cases involved major commercial chatbots repeating pro-Russian narratives. Further testing in January 2026 showed chatbots repeating false stories in half of the tested cases. The reliability of responses varies by language, with less common languages showing higher susceptibility to misinformation, likely due to English being a foundational language for AI models and weaker fact-checking ecosystems. Evidence suggests these narratives might be intentionally crafted for bots, with poor quality and low human engagement on some propaganda sites. Generative Engine Optimisation (GEO) is also emerging as a new tactic to influence AI-generated summaries.