05/12/2026
NewsGuard’s Reliability Ratings Now Appear in ChatGPT and Gemini Responses — Displaying the Trustworthiness of AI Chatbots’ News Sources
Trust Tool Now Includes Information About the Reliability of Sources AI Chatbots Cite in Responses
New York, NY — NewsGuard today announced that it has upgraded its browser extension, (available for download here) to work inside ChatGPT and Google Gemini — displaying its nonpartisan trust scores and providing access to its “Nutrition Labels” next to every source those AI tools cite.
The upgrade enables users to see, at a glance, whether the websites on which AI-generated answers are based are reliable or not — bringing NewsGuard’s nonpartisan, journalist-reviewed credibility data into the AI tools that hundreds of millions of people now consult for everyday research, news, and decision-making.
With this integration, subscribers to NewsGuard’s reliability ratings see trust scores displayed inline next to each source cited in ChatGPT and Gemini responses. For example, a chatbot answer that relies on TASS — the Russian government-owned wire service that NewsGuard rates 15 out of 100 including for publishing false or egregiously misleading Kremlin propaganda — will flag that score right in the chat window, warning AI users to “proceed with maximum caution.”
When ChatGPT cited TASS to answer a question about Ukrainian border crossings, NewsGuard flagged the Russian state outlet’s 15/100 trust score — and the specific journalistic standards violations behind it — directly in the chat window.
The same goes for known hoax sites, AI-generated content farms, and publishers flagged for health misinformation or election falsehoods. Credible sources get their scores surfaced too, giving users a clear, immediate signal about what they’re reading and where it came from.
The launch addresses a problem with the AI models that NewsGuard has documented since the public release of the first AI models. Since 2024, NewsGuard’s regularly published AI False Claims Monitor has tracked how the 10 leading AI chatbots respond to prompts involving false claims — testing whether they repeat the false claims, refuse to engage, or correctly debunk falsehoods. The results are alarming: as AI companies have pushed chatbots to be more responsive and enabled real-time web search, the top chatbots repeat false claims, including those from Russian, Chinese, or Iranian influence operations, 28.79 percent of the time when prompted on controversial news topics.
The core vulnerability, NewsGuard’s researchers found, is sourcing: Chatbots with web access are increasingly relying for their responses on unreliable publishers, low-quality content farms, and even state-controlled propaganda outlets, often presenting them alongside credible sources with no way for the user to tell the difference.
“AI tools are now one of the primary ways people get information — and they cite sources. But citing a source is only useful if you know whether to trust it,” said Matt Skibinski, NewsGuard’s Chief Operating Officer. “When a chatbot draws on a Russian state propaganda outlet or a site that’s been flagged for publishing health hoaxes, users have no way to know that — unless NewsGuard tells them. That’s exactly what this integration does.”
NewsGuard’s work is conducted by a team of journalists trained as information reliability analysts using a transparent and apolitical approach and a multi-layered review process to ensure high information quality.
NewsGuard’s browser extension also continues to provide its capabilities for everyday browsing, including trust scores and detailed Nutrition Labels for more than 13,000 news and information sites. The scores are displayed next to links across the web, including within search results and social media feeds. The extension is available on all leading browsers for $4.95/month, with a free two-week trial for new users.
About NewsGuard
NewsGuard helps consumers and enterprises find reliable information online with transparent and apolitical data and tools. Founded by media entrepreneur and award-winning journalist Steven Brill and former Wall Street Journal publisher Gordon Crovitz, NewsGuard’s products are all produced using a combination of human analysis scaled by technology.
NewsGuard’s detailed Source Reliability Ratings enable enterprises and consumers to identify reliable sources of information at scale, with coverage of more than 35,000 online sources accounting for 95%+ of engagement. Our continuously updated False Claim Fingerprints datastream helps clients identify and mitigate unreliable information, with data and analysis covering 60,000+ instances of false claims spreading online with detailed and precise data seeds built for automated tracking. Altogether, NewsGuard has collected more than 6.9 million information reliability data points for its clients and customers since its 2018 launch.