Popular News App Caught Publishing Completely Untrue AI Articles


NewsBreak — the most downloaded news app in the US — has been publishing made-up news articles generated by AI, a Reuters investigation reveals.

The app, helmed by CEO Jeff Zheng, is a news aggregator that syndicates licensed news from a range of national and international publishers. Lately, it's been using AI to rewrite and republish web-scraped local news and press releases. But according to Reuters, NewsBreak's AI efforts have been riddled with consequential missteps: since 2021, the service has reportedly published at least 40 inaccurate AI-generated news stories, many published under made-up bylines.

Many of these fake AI articles had real-world impact. Per the report, multiple NewsBreak posts repeatedly and inaccurately stated that a Colorado-based food kitchen was distributing food at times when it wasn't, resulting in the kitchen having to refuse hungry community members. A similar issue occurred in Pennsylvania, where AI-generated NewsBreak articles incorrectly claimed that a local charity would be hosting a 24-hour foot-care clinic for homeless people.

Do Harm

"You are doing HARM by publishing this misinformation," the charity wrote NewsBreak in a January email, according to Reuters, adding that "homeless people will walk to these venues to attend a clinic that is not happening."

In yet another instance flagged by Reuters, NewsBreak published an AI-spun article about an alleged Christmas Day shooting in New Jersey that turned out to be entirely fake, forcing local police to issue a statement.

"Nothing even similar to this story occurred on or around Christmas, or even in recent memory for the area they described," read the police response, according to Reuters. "It seems this 'news' outlet's AI writes fiction they have no problem publishing to readers."

All five of these erroneous articles were reportedly published under nonexistent bylines, further confusing the flow of already bad information — not to mention undermining the integrity and trustworthiness of the platform as a whole.

Alarm Bells

The issue also didn't go unnoticed internally.

In a May 2022 memo to Zheng, Norm Pearlstine — a former executive editor at the Wall Street Journal and the Los Angeles Times who was consulting for NewsBreak at the time — called attention to NewsBreak's dubious AI practices, telling the CEO that he could not "think of a think of a faster way to destroy the NewsBreak brand," according to Reuters.

But NewsBreak didn't halt its use of the AI tool. And in March of this year, per Reuters, it caveated in a homepage disclaimer that its articles "may not always be error-free."

It's a bad look for the service, which prides itself on being the "nation's leading local news app" in an era when local news is on a rapid and concerning decline. And sadly, it's only the latest aggregator of news and information to wreak havoc on its core product with AI.

More on AI and media: Meet AdVon, the AI-Powered Content Monster Infecting the Media Industry