CNET Pauses AI-Written Articles to Let Backlash Die Down

Stock image of robot hand typing on laptop
Stock image of robot hand typing on laptop


CNET’s AI is out of a job at the moment.

CNET told staff it would halt the publication of articles generated via artificial intelligence, in a Friday call, according to a report from The Verge. Or, at least, the company said it would pause the AI-article practice “for now,” as it waits out a stream of media criticism.

“I just want to reassure everybody: this will pass,” Lindsey Turrentine, a CNET executive, told staff regarding the recent coverage of the site’s AI use. “It’s uncomfortable, we will get through it, the news cycle will move on,” she added, in the hour-long call.

Read more

The storied tech media outlet, acquired in 2020 by private equity group Red Ventures, had been quietly publishing AI-generated content for months, under the byline “CNET Money Staff” or “CNET Money.” Other sites owned by Red Ventures, including Bankrate and CreditCards.com, have also been using the same, unnamed proprietary AI tool to generate content, Lance Davis, a Red Ventures executive, noted in the staff call, according to The Verge.

In the CNET stories, a small, hard-to-locate disclaimer on the “author” page was the only initial hint that a human hadn’t crafted the prose behind the posts. Until the true nature of those articles—almost entirely basic financial explainers transparently intended to garner clicks through SEO—was first widely noted in a January report from Futurism. And from there, multiple other major media outlets picked up the story.

A follow-up report from Futurism noted an alarming number of errors in just one of the AI-assisted posts. The article, titled “What is Compound Interest?,” misrepresented compound interest multiple times. CNET was forced to issue a hefty correction, and began reviewing the accuracy of all of its similarly produced content

Then, interviews with current and former CNET staff members revealed that the “CNET Money” posts were just the tip of the AI-iceberg. CNET had been using some form of automated assistance in posts for at least 18-months, and editorial staff were often left in the dark about when are where automation was being used, according to an earlier report from The Verge.

In the Friday staff call, Davis, Turrentine, and CNET editor-in-chief Connie Guglielmo offered staff more details about how AI was being used on the site. Editors using the AI tool can choose domains and sections for articles, pull data, and generate text. They can also include or add their own, human-written text or reporting.

Worth noting: multiple major news outlets use automation or AI for some content on their sites. For instance, the Associated Press began using AI to generate articles on earnings reports in 2014, and later expanded that to sports recaps. The Washington Post has used artificial intelligence to keep tabs on election results and football games.

However, CNET’s application of AI has been slightly different. By publishing a slew of articles with basic questions as headlines, meant to rank as high as possible in Google’s algorithmic assessment, the company was clearly trying to corner the click-market on financial advice—not free up journalists for other work. And all this from the same outlet that published an (actual, reported) article arguing that large language models wouldn’t replace humans in media. CNET didn’t immediately respond to Gizmodo’s request for comment.

Of course CNET and Red Ventures aren’t solely to blame in the rise of AI-written content on the web. Google has allowed these posts to flourish on its search results pages, providing companies ample financial incentive to continue pursuing artificial intelligence as an avenue of content creation.

“Our ranking team focuses on the usefulness of content, rather than how the content is produced,” Danny Sullivan, Google’s public search liaison, told Futurism. Though “useful” is a misleading term when SEO headlines can mask faulty financial advice. And, if Google continues to take that stance, the trustworthiness of your search results could get even shakier, as spammers and click-farmers aim to take advantage. Gizmodo reached out to Google with questions, but did not immediately receive a response.

More from Gizmodo

Sign up for Gizmodo's Newsletter. For the latest news, Facebook, Twitter and Instagram.

Click here to read the full article.