A new generation of content farms are harnessing AI to spin out clickbait — and they’re getting help from Google.

Last year, Icelandic teacher María Hjálmtýsdóttir wrote a column for The Guardian on the country’s experiment with a 36-hour workweek. The piece offered rich personal anecdotes that only a local could provide. Readers learned, for instance, that Hjálmtýsdóttir’s husband is using some of his newfound free time to chat with his fellow hobbyist pigeon keepers.

In the months since her Guardian piece came out, Hjálmtýsdóttir’s essay has been stripped of its color, repackaged, and republished at least a dozen times by “news outlets” that almost nobody has ever heard of.

“Iceland switched to a 4-day workweek — Gen Z was right all along,” stated a July 3 headline on Dixie Sun News, on a URL that once hosted a college newspaper. “Iceland embraced the 4-day workweek in 2019: 6 years later, Gen Z’s vision has been realized,” stated the Carroll County Observer, a former Maryland news site turned clickbait slop shop. “Iceland embraced a 4-day workweek in 2019 – Now, nearly six years on, all Gen Z forecasts have materialized,” read the headline on WECB.fm, a site falsely claiming to represent an Emerson College radio station.

These sites appear to be part of a new wave of AI-generated content farms that swoop in to seize dormant domains. Some of the AI news sites led previous lives unrelated to news, like Boston Organics, the website of a former produce delivery service that now covers everything from octopuses in British waters (“England is facing an unprecedented invasion, the problem is, it’s octopuses, and they’re devouring everything in their path”) to how long chili stays good in the fridge. In other cases, AI news articles are buried out of view of the homepage. Users who visit Paris2018.com — a site created for that year’s Gay Games — see no indication that it contains a plethora of AI-generated articles.

A sampling of stories on the Gay Games website

For readers, the sloppified news churned out by many of the sites is a reason to be wary of a confusing media landscape. And for beleaguered news publishers, the AI news sites threaten their investments in journalism, since the sites can almost instantly churn out versions of original reporting and capture valuable web traffic. Some of the sites get a boost from Google Discover, where I found many of them in my feed.

For more than half a century, staff at the Farmingdale Observer covered high school sports and local arts festivals in the Long Island town. In 2022, the paper’s owner at the time, Anton Media Group, merged it with the nearby Nassau Observer. Then, in May 2024, the Observer’s former website — which had been frozen on the same homepage stories since the merger — abruptly pivoted to buzzy viral content, from the latest in military tech to the must-have furniture color of 2025. The articles began showing up in Google Discover, where I first encountered them. (Disclosure: I worked for Google as a content strategist from 2013 to 2015.)

The site is a sham facsimile of the original. Joshua Schneps, CEO of Schneps Media, which owns the Farmingdale Observer, told me that when Anton Media Group accidentally let the domain lapse, an unknown party scooped it up.

The Farmingdale Observer published at least two separate ripoffs of Hjálmtýsdóttir’s story in May 2025, under slightly different headlines and by two different authors: “Rose Dixon” and “Bob Rubila.” Both include a fabricated quote that’s similar to the deck of Hjálmtýsdóttir’s original Guardian piece:

María Hjálmtýsdóttir, activist and teacher, tells us: “The shorter working week has been a great success in Iceland and has changed my family’s life. For 90% of Icelanders, the 36-hour week means less stress, more job satisfaction and more time to enjoy life.”

Dixon’s stories show remarkable breadth. Recent headlines include “This IKEA lamp looks ten times more expensive than it is, and it’s bound to sell out fast,” “China has reportedly launched the world’s first nuclear reactor that doesn’t run on uranium,” and “Few people know this but putting a coin in the freezer is one of the best tricks to avoid serious problems.“

“Dixon” reliably published two to three stories every day from January 28 to June 27. “Bob Rubila” and another byline, “Dave G. Rub,” also published frequently; their names may both be homages to a former reporter at the actual Long Island newspaper named Dave Gil de Rubio, who told me he hasn’t worked at the paper since 2022.

A sampling of stories by “Rose Dixon” on the Farmingdale Observer website

The Observer is emblematic of the new crop of seemingly AI-generated content farms sprouting from forgotten corners of the internet. They include Glass Almanac, a science and technology site that lists a fake address in San Francisco (“100 Tech Way”), pumps out sensationalized YouTube videos, and publishes articles under suspiciously generic author names (the photo of “Daniel Martinez” matches that of a French architect with a different name). Some articles include footnotes, a telltale sign that they were generated by tools like Perplexity. Matt McGee, who launched Glass Almanac in 2013 as a blog focused on Google Glass, told me in an email that he sold the long-dormant domain last year to unknown buyers in Kazakhstan.

AI-generated news sites have proliferated in the last two years. One study published by NewsGuard in May tallied nearly 1,300 AI-generated news sites across 16 languages. Some sites churn out AI-generated local news; Oregon Public Broadcasting identified one that published stories under identities stolen from real reporters. AI also represents a threat to any site that relies on freelancers; Business Insider and Wired recently retracted stories by a fictitious journalist seemingly written using AI.

The Farmingdale Observer abruptly stopped publishing new stories at the end of June, when an investigation by The Washington Examiner highlighted the outlet’s propensity for hyping Chinese military and technology. It’s not alone. Glass Almanac has repeatedly published stories praising the might of Chinese warships, claiming they have the U.S. on “high alert,” threaten to “reshape global sea power,” and “spark a new global power race.”

A site called African in Space — another repurposed domain with stories that showed up in my Google Discover feed earlier this month — has run articles touting the superiority of Chinese submarine defense systems, satellites, and wind turbines. The fake Emerson College radio website WECB.fm, which published a version of Hjálmtýsdóttir’s piece, has run stories praising Chinese steel production, battery innovation, and “fusion laser” development.

A sampling of stories on the African in Space website

It’s not clear who is behind the sites. But one of the pages offers some clues.

When I emailed the address listed on WECB.fm, a representative of Tremplin-Numerique, a digital marketing firm registered in Estonia, responded.

Instead of answering my questions, the representative, “Cyrielle,” sent a template email intended for advertisers. It offered to place sponsored articles on a variety of websites, from the Senegalese news site Seneweb to a Japanese site called Dog Magazine. Cyrielle’s spreadsheet also includes WECB.fm; the agency claims it drew more than 4 million visitors in the month of May and was listed on Google News. Prices for the sponsored articles ranged from €10 to €2,000 depending on the amount of traffic the sites received.

Cyrielle offered to write paid content for advertisers, charging between €35 and €75 (it was unclear if those stories would be generated using AI) and warned that content that linked out to “sensitive” advertisers — “casino, poker, crypto, CBD, dating, trading, VPN” — would cost two to three times more than the listed prices.

On its website, Tremplin-Numerique claims to have created 1,500 websites and generated €2 million in revenue using the AI content creation tool WP Auto.

It’s not clear which sites in Cyrielle’s spreadsheet are managed directly by Tremplin-Numerique or how many are powered by AI. A representative for Seneweb told me in an email that they do not accept sponsored content. Cyrielle did not respond to follow-up questions, and Romain Rafecas, listed on LinkedIn as Tremplin-Numerique’s cofounder, did not respond to my LinkedIn message.

It’s never been easier to launch an AI-powered news site. The internet is stocked with end-to-end tools that automate the entire process of creating and posting a news story, including text, photos, headlines, and social media posts.

Some, like Gravity Write, allow users to choose the language, tweak the formality of posts, and suggest topics for posts and blogs. Others, like WP Auto, can generate and translate content based on source websites selected by users, including traditional news sites. When the source sites add a new story or blog post, the AI sites can automatically create their own article.

The content on the websites sometimes surfaces on Google News and Google Discover, neither of which explicitly bars AI-generated content. “If you can appear in Google Discover or Google News, you’ve hit the jackpot,” said Matt Zimmerman, an entrepreneur and inventor of a content tool called Zimmwriter, aimed at marketers (“The AI writer trusted by SEO experts”) that uses generative AI startup Perplexity to create hundreds or thousands of articles. “You’re going to get 10,000, 50,000, 100,000 visitors in a single day.”

Some of those hits come at the expense of more traditional websites that run on human labor. Jake Ward, an entrepreneur who runs a content creation tool called byword.ai, described pulling off a “heist that stole 3.6 [million] total traffic from a competitor” by scraping a list of its articles and recreating versions of the stories using AI.

A spokesperson for Google didn’t respond to specific questions about how the company evaluates AI-generated news content, but noted that its spam policies include bans on expired domain abuse — taking over unrelated domains with the goal of manipulating search rankings.

“Our spam policies prohibit deceptive practices designed to manipulate Search or Discover,” the spokesperson said. “While we don’t comment on actions taken against individual sites, when we identify violations of our policies, we take action, which may include manual removal.”

Maria Hjálmtýsdóttir, the Icelandic author of the original Guardian piece, was unaware of the traction her piece generated on AI news sites until I contacted her.

Her initial reaction was to be flattered at the attention. “Then when I stop to think about it, it’s kind of scary,” she said in an interview. “You can’t trust anything.” As a social and gender studies teacher of high schoolers, Hjálmtýsdóttir was developing her own eye for AI writing.

“Many of our teachers are going back to just pen and paper and oral exams,” she said. “We don’t trust anything they hand to us digitally anymore.”

Ben Paviour is a freelance reporter based in Richmond, Va. He previously served as a local investigations fellow at the New York Times, covered Virginia state politics for local and national audiences on NPR and reported internationally in Phnom Penh, Cambodia.