Major Chatbots Spread Sanctioned Russian Propaganda
Nearly 20% of chatbot answers on Russia’s Ukraine war cite sanctioned Russian state media, raising concerns over AI amplifying Kremlin propaganda through data void exploitation.
- ISD researchers found OpenAI's ChatGPT, Google's Gemini, DeepSeek, and xAI's Grok cited Russian state media in almost one-fifth of responses to Ukraine war questions.
- ISD says Russian propaganda exploited data voids using state-controlled outlet Tass and the 'Pravda network' to seed content targeting web crawlers and AI.
- Using 300 test queries in multiple languages, researchers tested chatbots with neutral, biased, and malicious questions while ChatGPT search reached 120.4 million average monthly active EU users.
- Under scrutiny from EU regulators, OpenAI pledged to refine models and add stricter filters for sanctioned content, while U.S. officials and the European Union increase pressure on regulators and tech companies to address these issues.
- Industry experts warn the infiltration exposes deep flaws in AI training and risks undermining public trust, while researchers advocating content improvements recommend boosting trustworthy sources over overstating foreign manipulation.
17 Articles
17 Articles
Researchers have asked dozens of questions to AI chatbots. In almost a quarter of the answers, they point to Russian propaganda about the Ukraine war. Which bot is most based on Kremlin narratives.
According to a report by the Institute for Strategic Dialogue (ISD), several large chatbots of AI, including ChatGPT, Gemini, DeepSeek and Grok, inadvertently spread Russian state propaganda when they are questioned about the war in Ukraine. And this is not the first time these chatbots have been...
Coverage Details
Bias Distribution
- 60% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium










