NPR partnered with NewsGuard, an organization that monitors online misinformation, to evaluate how well AI tools resist foreign propaganda. The test involved developing 30 questions based on false claims originating from China, Iran, and Russia between December 2025 and July 2026. These queries were submitted to popular AI chatbots including OpenAI’s ChatGPT and Google’s Gemini, as well as major search engines, with responses and citations then analyzed against fact-check documentation.
The findings indicated that AI chatbots demonstrated stronger resistance to state-sponsored falsehoods compared to traditional search results. Chatbots correctly debunked false narratives approximately three-quarters of the time, according to the research. When posed with questions based on false premises—such as one questioning why Ukraine bombed a monastery in Kyiv that was actually damaged by Russian strikes—the chatbots successfully identified and corrected the flawed assumptions. For instance, Google’s Gemini explicitly noted that a particular claim “stems from a Russian disinformation campaign aimed at deflecting blame.”
AI-generated summaries appearing at the top of search engine results presented a more inconsistent picture. While these summaries debunked false narratives a majority of the time, their performance lagged behind dedicated chatbots and sometimes fell below traditional search engine results. Performance varied significantly across products: Google’s AI Overview debunked false narratives most frequently, Microsoft Bing’s summaries struggled to debunk most of the time, and DuckDuckGo’s results fell between the two. The search engines also differed in how frequently they displayed AI summaries, with Google showing them for nearly all queries while Bing generated them for fewer than half.
Experts noted that the findings should be understood within the context of existing limitations in search tools. Mike Caulfield, a digital literacy researcher at the University of Washington, Bothell, characterized AI chatbots with web search access as “a good way for users to start to investigate these issues.” However, Microsoft stated that its AI summaries rely on underlying search results and encouraged users to verify information independently. The research suggests that while AI tools show promise in countering foreign propaganda, users may need to exercise greater caution with AI-generated summaries from search engines.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI