USA Today Using AI to Generate Parts of Articles, Leaked Memo Shows

Gannett's AI optimism persists.

Key Points

The American newspaper giant Gannett is using AI to generate parts of its articles, The Verge reports.

A leaked memo reviewed by the Verge reveals that the media conglomerate — which owns the national title USA Today, in addition to hundreds of local newspapers — has rolled out a program that drafts AI-generated "Key Points," or a short AI-generated list of bulleted takeaways pinned to the top of a given article.

The Gannett memo states that the AI tool was trained by Gannett during a nine-month period, according to the Verge's reporting. The automated roundup tool "aims to enhance the reporting process and elevate the audience experience," according to the memo. How such a tool would do much of either, though — unless elevating the "audience experience" simply means keeping readers from having to read an article in full — is unclear.

Already Out There

Per the Verge, the new AI tool already appears to be in use. Articles that have employed the program are outfitted with a clear disclaimer, which reads that the "Key Points at the top of this article were created with the assistance of Artificial Intelligence (AI) and reviewed by a journalist before publication."

"No other parts of the article were generated using AI," the disclaimer adds. A quick Google search for the disclosure yields a long list of articles published by Gannett papers outfitted with the AI-drafted bullet points, including many in USA Today.

For its part, Gannett simply told the Verge that the memo "speaks for itself."

GAInett

This isn't the first time this week that Gannett has been in the news for its AI efforts. As Digiday reported earlier this week, Gannett employees at the Rochester, New York-serving Democrat & Chronicle newspaper are furious after their owner quietly altered contract language to say that AI can be used by newsrooms to "generate news content" with seemingly little to no boundaries.

AI has also caused credibility problems for the publisher. Last August, the newspaper giant was found publishing bizarre and error-laden AI-generated posts about local high school sports matchups in many of its regional papers. Not too long after that, writers at the USA Today site Reviewed publicly accused Gannett of publishing AI-generated articles bylined by fake authors, allegations that a recent Futurism investigation into the articles' provider further corroborated. (Gannett has denied that AI was used to generate the articles' text.)

In short, Gannett is leading the way to incorporate AI into journalism, albeit with sometimes scandalous results. But maybe the media giant's insistence on using AI says more about the state of the American newspaper economy than it does about the technology itself.

More on AI and journalism: Publishers Horrified at New Google AI Feature That Could Kill What's Left of Journalism