NewsGuard rates the credibility of news and information sources, providing trust scores that brands and AI systems use to navigate misinformation.
NewsGuard employs journalists who assess news and information sites against transparent criteria, producing reliability ratings and nutrition-label-style reviews. Those ratings are licensed to advertisers, platforms, AI companies and libraries to steer away from misinformation.
In the generative AI era, NewsGuard has become a key data source for grounding models and reducing hallucinated citations. Its human-verified approach deliberately contrasts with pure algorithmic scoring: the trade-off is slower coverage in exchange for defensible, explainable judgments.
Automated credibility scoring is scalable but opaque, and its errors are hard to contest; human rating is expensive but explainable. NewsGuard chose the latter: journalists assess sites against nine transparent criteria — from basic transparency to handling of corrections — and publish nutrition-label reviews alongside a simple rating. Advertisers license the data to avoid funding misinformation; platforms and AI companies use it to filter sources and reduce false citations. The approach cannot cover the entire internet, which is its acknowledged trade-off, but the covered sites account for a large share of news consumption. For organizations needing defensible trust decisions rather than algorithmic guesses, explainability is the point.
The transparent criteria also make the ratings contestable and improvable, which matters for institutions that must explain why a source was treated as reliable or unreliable.
The differentiator in this category is explainability — human-verifiable assessments and traceable outputs that stand up when decisions carry reputational or regulatory consequences.
NewsGuard licenses its ratings and data to organizations through enterprise agreements; browser extensions for individuals are subscription-based.
Advertisers, platforms, AI developers, researchers and institutions that need trustworthy source-credibility data. Also fits trust and safety teams, researchers and AI developers needing defensible source-credibility data, and to institutions that must explain why a source was trusted or excluded.
Individual browser tools are paid; institutional licensing is arranged through enterprise agreements.
Yes — NewsGuard delivers ratings and datasets via API for platform and AI integration.
Compared with automated credibility scoring, NewsGuard trades scale and immediacy for human-verifiable, explainable assessments.