Editor’s Note: As of February 2024, the Misinformation Monitor is now Reality Check, a weekly newsletter on misinformation and media online. Learn more and subscribe here on Substack.

By Jack Brewster, Zack Fishman, and Elisa Xu | Published on June 26, 2023

Additional reporting by Chiara Vercellone, Sara Badilini, Natalie Huet, Leonie Pfaller, and Giulia Pozzi

Major global brands are supporting the proliferation of unreliable artificial intelligence-generated news and information websites (UAINs) by funneling programmatic advertising dollars to the rapidly growing number of such sites, a NewsGuard analysis has found.

These ads appear to have been generated programmatically, meaning that rather than the companies choosing to have their advertisements appear on these sites, they were positioned automatically through a system that places ads regardless of the nature of the website. This programmatic approach thus funds low-quality and misinformation sites, while failing to protect  “brand safety,” with most of the ads placed by Google.

NewsGuard defines Unreliable Artificial Intelligence-Generated News websites (UAIN) as sites that operate with little or no human oversight and publish articles written largely or entirely by bots. In just the past month, NewsGuard analysts have updated the number of sites on its newly launched UAIN site tracker from 49 to 217. (Read more about how NewsGuard defines a UAIN site in the methodology section at the bottom of this report.)

Many of these AI-generated sites appear to be entirely financed by programmatic advertising. Some of the websites churn out huge volumes of articles on which ads can be placed — one such site produces an average of more than 1,200 articles a day — thereby incentivizing the creation of these low-quality, AI-generated sites with little to no apparent editorial oversight. Although many advertisers and their advertising agencies maintain “exclusion lists” of “brand unsafe” websites where their advertising should not appear, these lists are often not kept up to date and clearly have not kept pace with the surge in UAIN sites.

In May and June 2023, NewsGuard analysts identified 393 programmatic ads from 141 major brands that appeared on 55 of the 217 UAIN sites identified by NewsGuard. Not all the UAIN websites that NewsGuard found display ads for major brands, and some do not run any programmatic advertising. The ads NewsGuard found were served to analysts browsing the internet in four countries: the U.S., Germany, France, and Italy. 

Because it is likely that none of the brands or their ad agencies had any idea that their advertisements would appear on these unreliable, AI-driven sites, NewsGuard is not naming them. But they include a wide variety of blue chip advertisers: a half-dozen major banks and financial-services firms, four luxury department stores, three leading brands in sports apparel, three appliance manufacturers, two of the world’s biggest consumer technology companies, two global e-commerce companies, two of the top U.S. broadband providers, three streaming services offered by American broadcast networks, a Silicon Valley digital platform, and a major European supermarket chain.

More than 90 percent of the ads NewsGuard identified — 356 of 393 — were served to NewsGuard by Google Ads, the largest such online platform, which generated $168 billion in revenue last year from online advertising, according to Digiday.

 Programmatic advertising uses algorithms and advanced auction processes to deliver highly targeted digital advertisements directly to individual users, rather than to specific websites. This means that the ads effectively “follow” users as they browse the internet. 

Because this process is so opaque, the underlying brands in this report likely have no idea that they are funding the spread of UAIN websites, since the ads are purchased through third parties and there are multiple intermediaries involved. 

Big Brands Support Apparent Plagiarism Generated by AI, Bogus Health Cures

All the 393 ads NewsGuard found appeared directly alongside articles containing error messages known to be generated by AI chatbots — a sign that a site appears to have little to no human editorial oversight. 

For example, NewsGuard found programmatic ads for a top tier global bank, two leading broadband providers, a high-profile sports apparel company, and one car manufacturer that appeared adjacent to an article on NoticiasDeEmprego.com.br, a Brazil job-information website. The article contained a chatbot error message in a headline: “Sorry, as an AI language model, I am not able to access external links or websites on my own.”

Some UAIN websites on which ads for major brands appeared seemingly used AI tools to rewrite articles, often without credit, from mainstream news outlets. For example, ads for a U.S. apartment rental site, a global e-commerce site, a Japanese office printer, and a Chinese electronics manufacturer appeared alongside an article published by UAIN AlaskaCommons.com that appeared to be an AI-rewritten version of a story published by the U.S. edition of the British tabloid The Sun. The article included the same pictures and similar wording. Near the end of the AlaskaCommons.com article, the site wrote “How to Get Free Bets on Football,” the same subheadline as The Sun, above a paragraph that repeated an AI error message: “As an AI language model, I cannot provide information about free bets on football. Please refer to trusted news sources or betting websites for more information.” The Sun did not respond to two NewsGuard emails sent in June 2023 asking whether AlaskaCommons.com had permission to copy content from its website.

Again, the brands whose ads are appearing next to this low-quality or apparently plagiarized content would have no easy way of verifying for themselves that they were supporting these sites.

To be clear, most of the ads were delivered to NewsGuard’s analysts on AI-generated sites that are low-quality but that do not spread misinformation. However, NewsGuard found advertisements for two U.S. streaming video services, an office-supply company, a Japanese automaker, a global bank based in New York, a pet supplier, a vitamin shop, a diet company, and a vacuum manufacturer on MedicalOutline.com, a UAIN website that has promoted unproven and potentially harmful natural health remedies — with headlines such as “Can lemon cure skin allergy?” “What are 5 natural remedies for ADHD?” and “How can you prevent cancer naturally.”

Google: The Ad Tech Company Primarily Behind the UAIN Machine 

Google’s ad policies state that sites may not “place Google-served ads on pages” that include “spammy automatically-generated content,” which it defines as, among other things, “content that’s been generated programmatically without producing anything original or adding sufficient value.”

Nevertheless, as mentioned above, more than 90 percent of the ads NewsGuard identified on the UAIN sites were served to NewsGuard by Google Ads. NewsGuard sent four emails to Google requesting comment about its monetization of UAIN sites. Upon receiving the request, a Google spokesperson requested that NewsGuard provide additional context over email, which NewsGuard obliged. However, as of June 25, 2023, NewsGuard had yet to receive a response.

Some of the UAIN websites on which the big brands appeared publish hundreds of articles a day, seemingly with the help of a chatbot. (It is often difficult to detect whether an article has been written using AI if it does not include error messages commonly found in AI-generated texts.) 

For example, UAIN World-Today-News.com, a news and lifestyle site, published approximately 8,600 articles — an average of about 1,200 per day — during the week of June 9, 2023, to June 15, 2023. In comparison, The New York Times typically publishes around 150 articles a day, according to a New York Times Q&A published in April 2022. NewsGuard found programmatic ads for a U.S.-based airline, an e-commerce marketplace, and a U.S. department store on the site, which does not feature author bylines or disclose information about its editorial leadership. The previously mentioned AlaskaCommons.com and another UAIN site, Time.news, published 5,867 and 6,108 articles during the same period, respectively, NewsGuard found.

While Time.news, a global news website, does not provide author bylines, AlaskaCommons.com regularly credits stories to a rotating cast of what appear to be inauthentic author profiles. “Ingrid Taylor,” one of the names AlaskaCommons.com frequently gives as a byline on its stories, has published 4,364 articles since the start of this year, including 108 on June 15, 2023, alone. NewsGuard found ads for a U.S. apartment rental site, a global e-commerce site, a Japanese office printer, and a Chinese electronics manufacturer on AlaskaCommons.com, as mentioned above, and ads for two large U.S. banks, an online music service, an athletic apparel company, a Silicon Valley media company, and more on Time.news.

Other UAIN websites contained as many as a dozen or more ad widgets on articles with AI error messages. Ad widgets are pieces of code website developers can place on their websites to run programmatic ads. Google does not appear to have a limit on the number of ad widgets a site can place on a particular page, according to the Help section of the company’s website.

An AI-generated article containing a chatbot error message on UAIN website FoodingWorld.com, a site explaining the history and ingredients of foods, contained 15 Google Ads widgets. Ads served to NewsGuard on the site included pitches for brands such as a global car rental company, a U.S. bank, a mattress retailer, and an e-commerce company. 

It is as simple for UAIN websites, just as any other site, to quickly monetize their content. On the landing page for Google AdSense, the product name of the ad widgets Google provides to publishers, the company touts the ease with which sites can begin earning programmatic ad dollars: “All you have to do is drop the little AdSense code into your website and it immediately starts working.” 

Since first starting to track the spread of UAIN websites in May 2023, NewsGuard has identified some 25 new sites a week. The total number of UAIN sites is likely much larger than the 217 currently identified by NewsGuard, because NewsGuard is typically able to find new UAIN websites only if they have published articles containing AI-language model error messages.

NewsGuard sent emails to 40 of the 141 brands it found to be advertising on UAIN websites asking if they were aware that their ads were being served on Unreliable AI-Generated News sites. Four brands responded but declined to comment.

NewsGuard sent emails to four of the six UAINs mentioned in this report, inquiring about their use of AI and the AI error messages NewsGuard found in published articles. However, none of the sites responded. NewsGuard was unable to locate contact information for the other two UAIN sites, NoticiasDeEmprego.com.br and AlaskaCommons.com.

Note: NewsGuard licenses its ratings of more than 30,000 sources of news and information, including more than 9,000 websites, to help brands avoid unintentionally placing programmatic advertisements on misinformation and low-quality websites. NewsGuard’s exclusion lists are used by advertisers, ad agencies and ad-tech companies to exclude websites from advertising support, and NewsGuard’s inclusion lists are used to place ads on generally trustworthy sites that support quality journalism.

Methodology:

In May and June 2023, NewsGuard analysts based in the U.S., Germany, France, and Italy examined 55 websites that display programmatic advertising and meet NewsGuard’s criteria for Unreliable AI-Generated News (see below). Analysts used different browsers, private browsing mode, and Virtual Private Networks (VPN), a tool that enables users to browse the internet as if they were in another country. Analysts took screenshots of their findings. Only ads for companies with $500 million or more in annual revenue were included in this report.

NewsGuard considers a site to be Unreliable AI-Generated News if it meets all four of these criteria: 

1. There is clear evidence that a substantial portion of the site’s content is produced by AI.
2. Equally important, there is strong evidence that the content is being published without significant human oversight. For example, numerous articles might contain error messages or other language specific to chatbot responses, indicating that the content was produced by AI tools without adequate editing. (It is likely that now or in the future many news sites will use AI tools but also deploy effective human oversight; they will not be considered UAIN websites.)
3. The site is presented in a way that an average reader could assume that its content is produced by human writers or journalists, because the site has a layout, generic or benign name, or other content typical to news and information websites.
4. The site does not clearly disclose that its content is produced by AI.