Report March 2026
Submitted
In June 2022, NewsGuard signed onto all the Code of Practice’s measures that are relevant to its services, committing to maintain its journalistic practices, which include an independent, unbiased, and apolitical approach; full transparency on its methodology; and providing publishers a ‘right to be heard’ by calling for comment when publishers are deemed to fail certain criteria. Furthermore, NewsGuard also committed to continue to strengthen its media-literacy efforts with public libraries and schools to help users develop their critical thinking and online awareness through its browser-extension tool.
Here is a summary of how we implemented our commitments in 2025.
1. EMPOWERING USERS
NewsGuard’s Reliability Ratings are based on nine apolitical and basic journalistic criteria that assess the credibility and transparency of a news or information site. Our nine criteria are applied equally to all news sites, regardless of their size or political orientation — from mainstream media outlets to small blogs — while allowing all sites to exercise their right to be heard.
Our criteria are basic journalistic principles and are inherently apolitical. They are also completely transparent and explained in great detail on our website in several languages spoken in the EU (namely English, French, German and Italian) along with the relative weighting of each criterion, depending on its importance.
Our rating process is designed to ensure our criteria are applied in an unbiased manner. After an analyst produces a first rating, with an associated Nutrition Label explaining in detail why he or she arrived at such a rating, the analysis is edited by at least two editors. Then, if any doubt remains about how to apply a specific criterion, the rating is discussed during a full staff meeting hosted by our two co-CEOs and co-Editors-in-Chief, Steven Brill and Gordon Crovitz. Over the years, this process has led NewsGuard to assign high ratings and poor ratings to sites of all political leanings.
To ensure fairness, publishers are also put at the center of our rating process and given a chance to comment whenever our team flags an issue with the credibility or transparency of their site. We believe this publisher right of reply is fundamental to fair and accurate ratings.
For the sake of accountability, readers can see the credentials and backgrounds of everyone responsible for every NewsGuard Reliability Rating and Nutrition Label that they read.
Because our tools are meant to empower users so that they can make informed decisions on which sources to trust and which to be wary of, we also allow users to submit suggestions of sources to rate, if we have not yet rated them, and to send feedback on existing ratings. All inbound messages are reviewed by our team.
Finally, to ensure complete independence, we accept no fees from news websites to rate them. Our revenue comes from license fees that platforms, ad agencies, brands, media monitoring companies, AI companies, government entities and researchers — among other groups — pay to use our data.
News consumers are able to access NewsGuard ratings through our licensees as well as directly from NewsGuard’s browser extension. Companies providing internet browsers, search, social media and other services make our ratings available to their users to empower them with information about the nature of news sources they see online. Individuals can also subscribe to our ratings through a browser extension and mobile version.
At a time when generative AI models often “hallucinate” and create and spread false claims about topics in the news, NewsGuard data also protects news consumers when they use large-language models, and provide product teams and trust & safety teams within generative AI companies with protections to ensure that their AI systems are reliable and that risks are mitigated. This protects news consumers from being confronted to false claims in AI responses to prompts and also prevents malign actors running disinformation operations, including for the governments of Russia, China and
Iran, from abusing AI models to spread their narratives and inundate the responses of Western chatbots.
In 2025, NewsGuard also provided its media literacy browser extension for free to approximately 270 public libraries in the EU (in Italy, France, Germany and Slovenia) helping thousands of patrons navigate the internet more safely and access additional information and context on the sources of news and information they encounter online.
From January 1, 2025 to December 31, 2025, NewsGuard participated in many media literacy seminars, awareness-raising events, and discussions with misinformation and disinformation experts in the following EU countries: Belgium, France, Germany, Ireland, Italy, Spain and Norway. These events reached a total of approximately 5,000 participants, including educators and librarians who in turn could reach hundreds of students and library users.
2. SUPPORTING TRUSTWORTHY JOURNALISM AND DEMONETIZING FALSE CLAIMS AND STATE-SPONSORED DISINFORMATION
NewsGuard’s brand safety service, BrandGuard, enables brands to invest in ad inventory on high quality news sites that publish trustworthy journalism — and avoid placing ads on websites that repeatedly publish false claims, state-sponsored propaganda, undisclosed AI-generated content, or unreliable news. Advertisers using BrandGuard can access NewsGuard’s ratings to target or avoid ad placements based on their brand values and safety guidelines, enabling each brand to tailor its approach to advertising on news while enforcing the publisher’s right to be heard.
BrandGuard is powered entirely by NewsGuard’s Reliability Ratings of news and information websites, which are compiled by a team of experienced journalists and editors based on nine apolitical journalistic criteria. Based on the criteria, each publisher receives an overall rating level ranging from “High Credibility” to “Credible with Exceptions” to “Proceed with Caution,” along with a 0-100 reliability score and an assessment on each of the nine criteria. Our rigorous rating process is explained in great detail on our website, on a page called “Website Rating Process and Criteria” (which is available in several EU languages: English, French, Italian and German.)
As the page describes, our process is transparent and accountable to everyone involved — including publishers, advertisers, and the general public. Each criterion is defined in significant detail, with numerous examples of how a publisher would pass or fail the criterion. Each site’s score is derived entirely from our assessments of those criteria, which are each assigned a specific number of points as outlined on our website. For each rating, we provide a detailed written Nutrition Label report that explains why NewsGuard made its determination on each of the criteria, provides evidence and examples to back up its assessments, and includes any relevant comments from the publisher.
Before publishing a rating or update, we always seek feedback from publishers that fail any of our criteria. We hold ourselves to the same standards we expect of the websites we rate, which means transparently conducting responsible journalism and allowing publishers to exercise their right to reply. Each assessment of a website is made public via NewsGuard’s browser extension, which is available for public subscription. NewsGuard subscribers and rated publishers can read NewsGuard’s detailed analyses, called “Nutrition Labels,” which explain why NewsGuard made its determination on each of the criteria, provide evidence and examples to back up its assessments, and include any relevant comments from the publisher. Additionally, we also invite publishers who disagree with our rating to provide feedback on a dedicated page of our website.
This transparent process is documented on NewsGuard’s website and in our Nutrition Labels, and allows publishers not only the right of reply — but also an opportunity to improve. More than 2,230 websites rated by NewsGuard have taken steps to improve editorial practices after being contacted by our team during the rating process — resulting in improved ratings and scores and, in some cases, leading advertisers to monetize their websites.
To ensure our process remains strictly apolitical, NewsGuard relies on apolitical criteria when rating a site (e.g. a corrections policy cannot be partisan, as there is no conservative or progressive way to regularly publish corrections) and carries out a manual and rigorous editing process involving approximately five journalists and editors per website rating, ensuring no rating is the assessment of a single person. The most debated analyses undergo a final review step in which it is shared with the full team of NewsGuard analysts, including the two co-CEOs, to raise any issues and ensure consistency.
3. EMPOWERING THE RESEARCH COMMUNITY
As a member of the Italian Digital Media Observatory since 2021, NewsGuard regularly publishes its content and analysis on disinformation in Italy and in Europe on IDMO’s portal, contributing to the consortium’s media literacy efforts.
NewsGuard has various partnerships and collaborations with research institutions and universities that study disinformation, such as La Sapienza University in Rome, Ca’ Foscari University in Venice, Carlo Bo University in Urbino, University of Salerno, the European University Institute in Florence, the Italian National Research Council, IMT School in Lucca, the Universität der Bundeswehr München and the German Max-Planck-Institute in Germany and the French National Research Institute for Digital Science and Technology or Inria in France.
In 2025, NewsGuard’s team produced regular newsletters called Reality Check covering false claims and disinformation online with exclusive data from nine countries including four Member States (France, Italy, Austria, and Germany). NewsGuard also published False Claims tracking centers on the year’s most notable news events, to monitor the false claims they generated, and document their origin and spread. These included an Israel-Iran War Tracker, German Elections Misinformation Tracker, and Southern California Wildfire Tracker.
Our previous tracking centers on Israel-Hamas War False Claims, Russia-Ukraine War Disinformation and on AI-enabled False Claims were also regularly updated throughout the year. These reports are distributed in all the countries NewsGuard operates in and are available for free on NewsGuard’s website in English, French, Italian, and German. NewsGuard also continues to publish a State-Sponsored Disinformation Risk Briefing focused on hostile information operations by the governments of Russia, China and Iran targeting citizens in the EU and other democracies.
NewsGuard also regularly publishes Special Reports on disinformation focusing on different topics, trends, and platforms (17 in total in 2025), in addition to audits of the top 11 chatbots, which we started publishing in July 2024 (8 in total in 2025, including one testing responses in French, English, German, Italian, and Spanish, among others.)
NewsGuard’s white paper, “Fighting Misinformation with Journalism, not Algorithms,” which is published on our website, outlines independent research on the effect of using human-curated news reliability ratings to mitigate false news, some of which has been conducted by leading academic institutions and other top scholars using NewsGuard’s Reliability Ratings dataset.
Here is a summary of how we implemented our commitments in 2025.
1. EMPOWERING USERS
NewsGuard’s Reliability Ratings are based on nine apolitical and basic journalistic criteria that assess the credibility and transparency of a news or information site. Our nine criteria are applied equally to all news sites, regardless of their size or political orientation — from mainstream media outlets to small blogs — while allowing all sites to exercise their right to be heard.
Our criteria are basic journalistic principles and are inherently apolitical. They are also completely transparent and explained in great detail on our website in several languages spoken in the EU (namely English, French, German and Italian) along with the relative weighting of each criterion, depending on its importance.
Our rating process is designed to ensure our criteria are applied in an unbiased manner. After an analyst produces a first rating, with an associated Nutrition Label explaining in detail why he or she arrived at such a rating, the analysis is edited by at least two editors. Then, if any doubt remains about how to apply a specific criterion, the rating is discussed during a full staff meeting hosted by our two co-CEOs and co-Editors-in-Chief, Steven Brill and Gordon Crovitz. Over the years, this process has led NewsGuard to assign high ratings and poor ratings to sites of all political leanings.
To ensure fairness, publishers are also put at the center of our rating process and given a chance to comment whenever our team flags an issue with the credibility or transparency of their site. We believe this publisher right of reply is fundamental to fair and accurate ratings.
For the sake of accountability, readers can see the credentials and backgrounds of everyone responsible for every NewsGuard Reliability Rating and Nutrition Label that they read.
Because our tools are meant to empower users so that they can make informed decisions on which sources to trust and which to be wary of, we also allow users to submit suggestions of sources to rate, if we have not yet rated them, and to send feedback on existing ratings. All inbound messages are reviewed by our team.
Finally, to ensure complete independence, we accept no fees from news websites to rate them. Our revenue comes from license fees that platforms, ad agencies, brands, media monitoring companies, AI companies, government entities and researchers — among other groups — pay to use our data.
News consumers are able to access NewsGuard ratings through our licensees as well as directly from NewsGuard’s browser extension. Companies providing internet browsers, search, social media and other services make our ratings available to their users to empower them with information about the nature of news sources they see online. Individuals can also subscribe to our ratings through a browser extension and mobile version.
At a time when generative AI models often “hallucinate” and create and spread false claims about topics in the news, NewsGuard data also protects news consumers when they use large-language models, and provide product teams and trust & safety teams within generative AI companies with protections to ensure that their AI systems are reliable and that risks are mitigated. This protects news consumers from being confronted to false claims in AI responses to prompts and also prevents malign actors running disinformation operations, including for the governments of Russia, China and
Iran, from abusing AI models to spread their narratives and inundate the responses of Western chatbots.
In 2025, NewsGuard also provided its media literacy browser extension for free to approximately 270 public libraries in the EU (in Italy, France, Germany and Slovenia) helping thousands of patrons navigate the internet more safely and access additional information and context on the sources of news and information they encounter online.
From January 1, 2025 to December 31, 2025, NewsGuard participated in many media literacy seminars, awareness-raising events, and discussions with misinformation and disinformation experts in the following EU countries: Belgium, France, Germany, Ireland, Italy, Spain and Norway. These events reached a total of approximately 5,000 participants, including educators and librarians who in turn could reach hundreds of students and library users.
2. SUPPORTING TRUSTWORTHY JOURNALISM AND DEMONETIZING FALSE CLAIMS AND STATE-SPONSORED DISINFORMATION
NewsGuard’s brand safety service, BrandGuard, enables brands to invest in ad inventory on high quality news sites that publish trustworthy journalism — and avoid placing ads on websites that repeatedly publish false claims, state-sponsored propaganda, undisclosed AI-generated content, or unreliable news. Advertisers using BrandGuard can access NewsGuard’s ratings to target or avoid ad placements based on their brand values and safety guidelines, enabling each brand to tailor its approach to advertising on news while enforcing the publisher’s right to be heard.
BrandGuard is powered entirely by NewsGuard’s Reliability Ratings of news and information websites, which are compiled by a team of experienced journalists and editors based on nine apolitical journalistic criteria. Based on the criteria, each publisher receives an overall rating level ranging from “High Credibility” to “Credible with Exceptions” to “Proceed with Caution,” along with a 0-100 reliability score and an assessment on each of the nine criteria. Our rigorous rating process is explained in great detail on our website, on a page called “Website Rating Process and Criteria” (which is available in several EU languages: English, French, Italian and German.)
As the page describes, our process is transparent and accountable to everyone involved — including publishers, advertisers, and the general public. Each criterion is defined in significant detail, with numerous examples of how a publisher would pass or fail the criterion. Each site’s score is derived entirely from our assessments of those criteria, which are each assigned a specific number of points as outlined on our website. For each rating, we provide a detailed written Nutrition Label report that explains why NewsGuard made its determination on each of the criteria, provides evidence and examples to back up its assessments, and includes any relevant comments from the publisher.
Before publishing a rating or update, we always seek feedback from publishers that fail any of our criteria. We hold ourselves to the same standards we expect of the websites we rate, which means transparently conducting responsible journalism and allowing publishers to exercise their right to reply. Each assessment of a website is made public via NewsGuard’s browser extension, which is available for public subscription. NewsGuard subscribers and rated publishers can read NewsGuard’s detailed analyses, called “Nutrition Labels,” which explain why NewsGuard made its determination on each of the criteria, provide evidence and examples to back up its assessments, and include any relevant comments from the publisher. Additionally, we also invite publishers who disagree with our rating to provide feedback on a dedicated page of our website.
This transparent process is documented on NewsGuard’s website and in our Nutrition Labels, and allows publishers not only the right of reply — but also an opportunity to improve. More than 2,230 websites rated by NewsGuard have taken steps to improve editorial practices after being contacted by our team during the rating process — resulting in improved ratings and scores and, in some cases, leading advertisers to monetize their websites.
To ensure our process remains strictly apolitical, NewsGuard relies on apolitical criteria when rating a site (e.g. a corrections policy cannot be partisan, as there is no conservative or progressive way to regularly publish corrections) and carries out a manual and rigorous editing process involving approximately five journalists and editors per website rating, ensuring no rating is the assessment of a single person. The most debated analyses undergo a final review step in which it is shared with the full team of NewsGuard analysts, including the two co-CEOs, to raise any issues and ensure consistency.
3. EMPOWERING THE RESEARCH COMMUNITY
As a member of the Italian Digital Media Observatory since 2021, NewsGuard regularly publishes its content and analysis on disinformation in Italy and in Europe on IDMO’s portal, contributing to the consortium’s media literacy efforts.
NewsGuard has various partnerships and collaborations with research institutions and universities that study disinformation, such as La Sapienza University in Rome, Ca’ Foscari University in Venice, Carlo Bo University in Urbino, University of Salerno, the European University Institute in Florence, the Italian National Research Council, IMT School in Lucca, the Universität der Bundeswehr München and the German Max-Planck-Institute in Germany and the French National Research Institute for Digital Science and Technology or Inria in France.
In 2025, NewsGuard’s team produced regular newsletters called Reality Check covering false claims and disinformation online with exclusive data from nine countries including four Member States (France, Italy, Austria, and Germany). NewsGuard also published False Claims tracking centers on the year’s most notable news events, to monitor the false claims they generated, and document their origin and spread. These included an Israel-Iran War Tracker, German Elections Misinformation Tracker, and Southern California Wildfire Tracker.
Our previous tracking centers on Israel-Hamas War False Claims, Russia-Ukraine War Disinformation and on AI-enabled False Claims were also regularly updated throughout the year. These reports are distributed in all the countries NewsGuard operates in and are available for free on NewsGuard’s website in English, French, Italian, and German. NewsGuard also continues to publish a State-Sponsored Disinformation Risk Briefing focused on hostile information operations by the governments of Russia, China and Iran targeting citizens in the EU and other democracies.
NewsGuard also regularly publishes Special Reports on disinformation focusing on different topics, trends, and platforms (17 in total in 2025), in addition to audits of the top 11 chatbots, which we started publishing in July 2024 (8 in total in 2025, including one testing responses in French, English, German, Italian, and Spanish, among others.)
NewsGuard’s white paper, “Fighting Misinformation with Journalism, not Algorithms,” which is published on our website, outlines independent research on the effect of using human-curated news reliability ratings to mitigate false news, some of which has been conducted by leading academic institutions and other top scholars using NewsGuard’s Reliability Ratings dataset.