{"data":{"id":2289,"lang":"en","title":"AI Chatbots Debunk Foreign Propaganda 75% of the Time, Beating Search Engines","summary":"NPR and NewsGuard found AI chatbots correctly debunked state propaganda from China, Iran, and Russia about 75% of the time, beating search-engine AI summaries.","category":"ai","category_label":"AI","published_at":"2026-08-31T18:02:27Z","updated_at":"2026-08-31T18:02:27Z","url":"https://techtools.cz/tools/news/article/2289-ai-chatbots-debunk-foreign-propaganda-75-of-the-time-beating-search-engines","api_url":"https://techtools.cz/news-api/v1/articles/2289?lang=en","markdown_url":"https://techtools.cz/news-api/v1/articles/2289.md?lang=en","image_url":"https://cdn.kpbs.org/dims4/default/281d258/2147483647/strip/true/crop/4871x3000+231+0/resize/880x542!/quality/90/?url=https%3A%2F%2Fnpr-brightspot.s3.amazonaws.com%2Fa8%2F06%2Fa0997e9e4141bb678bd7f1b92548%2F2026-07-ai-state-answers-final.jpg","word_count":181,"reading_time_minutes":1,"sources":[{"url":"https://www.npr.org/2026/08/30/nx-s1-5876436/chatbots-search-propaganda","domain":"npr.org"}],"translation":{"lang":"cs","id":2362,"title":"AI chatboti odhalí zahraniční propagandu v 75 % případů, poráží vyhledávače","url":"https://techtools.cz/tools/news/clanek/2362-ai-chatboti-odhali-zahranicni-propagandu-v-75-pripadu-porazi-vyhledavace","api_url":"https://techtools.cz/news-api/v1/articles/2362?lang=cs"},"content":"A joint investigation by NPR and NewsGuard found that popular AI chatbots correctly debunked state-sponsored propaganda from China, Iran, and Russia roughly 75% of the time, outperforming traditional search engines and AI-generated search summaries.\n\nResearchers built 30 test queries around 15 false narratives that circulated between December 2025 and July 2026, then posed them to ChatGPT, Google Gemini, Microsoft Copilot, Meta AI, Grok, and Claude. AI-generated summaries sitting atop search results fared worse: Google's AI Overview debunked most claims, Microsoft Bing's summaries failed to push back on a majority of them, and DuckDuckGo landed in between. Roughly 1 in 9 factual claims in Google's AI Overviews lacked any source support at all.\n\nDigital literacy expert Mike Caulfield called the chatbots' results strong by classroom standards, saying that if a teacher saw three-quarters of students answer similarly correctly, \"you would be ecstatic.\" The findings suggest that as more people turn to conversational AI over search for news and current events, standalone chatbots are currently a more reliable first line of defense against disinformation than the AI summaries increasingly baked into search results."}}