Researchers say a coordinated Russian-linked disinformation campaign is increasingly influencing responses generated by leading artificial intelligence chatbots, raising concerns that AI systems are becoming a new battlefield in information warfare. According to a report by the U.K.-based think tank Demos, large language models including ChatGPT, Gemini, Claude, Mistral, and Grok were found to repeat or discuss fabricated pro-Kremlin narratives after those claims had been deliberately seeded across networks of deceptive websites using a tactic known as Generative Engine Optimization (GEO). The findings suggest that hostile actors are shifting their focus from manipulating social media users directly to manipulating the information AI systems consume, highlighting the need for stronger safeguards, improved source verification, and greater cooperation among AI developers to protect the integrity of AI-generated information.
Sources
- https://www.thetimes.com/business/technology/article/ai-chatbots-russian-propaganda-ukraine-war-zv6lsc0hl
- https://www.euronews.com/my-europe/2026/07/27/ai-chatbots-citing-russian-propaganda-sourced-from-eu-sanctioned-outlet
- https://www.washingtonpost.com/technology/2025/04/17/llm-poisoning-grooming-chatbots-russia
Key Takeaways
- Coordinated foreign influence campaigns are evolving beyond social media by attempting to manipulate the information sources that AI chatbots rely upon when generating responses.
- Researchers found that some major AI systems repeated or inadequately challenged false Russian narratives, demonstrating that even advanced language models remain vulnerable to sophisticated information operations.
- AI developers face growing pressure to improve source evaluation, identify coordinated manipulation campaigns more rapidly, and collaborate across the industry to prevent adversarial actors from poisoning AI knowledge ecosystems.
In-Depth
The emergence of generative artificial intelligence has created an entirely new front in the global information war. Rather than focusing exclusively on persuading human audiences through social media, researchers now argue that sophisticated influence operations are increasingly aimed at the AI systems millions of people consult for news, background information, and explanations of world events. That represents a strategic shift with potentially long-term consequences.
According to the research, pro-Kremlin actors allegedly flooded the internet with fabricated stories designed less to attract readers than to be indexed by search engines and absorbed into the information ecosystem available to AI models. The objective is straightforward: if enough seemingly independent sources repeat the same false narrative, an AI system may eventually treat those claims as credible enough to mention or even repeat.
For conservatives who have long criticized institutional bias in technology and media, these findings underscore a broader lesson. AI systems are only as trustworthy as the information environment from which they learn. Whether manipulation comes from foreign governments, political activists, corporations, or other special interests, attempts to distort AI-generated information threaten the public’s ability to obtain accurate answers.
The challenge now extends beyond improving individual AI models. It requires identifying coordinated influence campaigns, strengthening source validation, and increasing transparency regarding how AI systems evaluate conflicting information. As AI becomes an increasingly common gateway to knowledge, protecting it from organized manipulation will become as important as defending traditional media and critical digital infrastructure from foreign interference.