POLISCOPE
Back to feed
NewsKPBS Public MediaAugust 30, 2026San Diego

Study finds AI chatbots outperform search engines in identifying foreign propaganda

AI chatbots successfully debunked foreign-sourced falsehoods and avoided spreading misinformation more effectively than traditional search engines. However, AI-generated summaries placed above search results performed less accurately than the chatbots themselves.

artificial intelligencemisinformationtechnologymedia literacy
Read the full story at KPBS Public Media

Why It Matters

This study highlights the varying reliability of AI tools in filtering misinformation, which impacts how residents consume news and verify information online.

Key Facts

  • AI chatbots demonstrated a higher success rate in debunking foreign propaganda compared to standard search engines.
  • AI chatbots were less likely to uncritically repeat or spread foreign-sourced falsehoods than search engines.
  • AI-generated summaries displayed at the top of search results performed worse at identifying misinformation than standalone AI chatbots.

Who's Mentioned

organizationNewsGuard“Provided fact-check documents and research support.”organizationOpenAI“Developer of ChatGPT.”otherAI chatbots“The primary technology being tested for propaganda detection.”personMike Caulfield“Digital literacy expert at the University of Washington, Bothell.”other30“Number of questions used in the experiment.”organizationDuckDuckGo“Search engine provider.”othersearch engines“The comparative baseline used in the study.”organizationGoogle“Developer of Gemini and AI Overview.”organizationMicrosoft“Developer of Bing AI services.”personMorgan Wack“Postdoctoral researcher at the University of Zurich.”organizationNPR“Conducted the experiment on AI and disinformation.”