Top News App Found Spreading False AI-Generated Stories

After the United States’ most downloaded local news app, NewsBreak, disseminated an AI-generated story about a fabricated New Jersey shooting last Christmas Eve, New Jersey police had to issue an online statement to reassure citizens that the story was “entirely false,” Reuters reported.

New Jersey authorities clarified on Facebook that “nothing even similar to this story occurred on or around Christmas, or even in recent memory for the area they described,” criticizing NewsBreak for using AI to publish fabricated stories.

Despite attracting over 50 million monthly users, NewsBreak took four days to remove the fake shooting story. According to Reuters, this incident was not isolated, with at least 40 misleading or erroneous stories published via its AI tool since 2021. The tool scrapes the web and assists in rewriting local news stories.

The impact of these misleading AI news stories has been tangible, seven former NewsBreak employees, speaking anonymously, informed Reuters. For instance, Colorado’s Food to Power food bank had to turn people away due to inaccurate information about distribution times posted by the app. In another case, Pennsylvania charity Harvest912 had to deny services to homeless individuals due to NewsBreak falsely promoting a 24-hour foot-care clinic.

NewsBreak claimed to Reuters that erroneous articles involving the two mentioned charities were removed but blamed the charities for supposedly posting inaccurate information on their own websites.

In a step aimed at reducing liability, NewsBreak added a disclaimer to its homepage in early March, warning readers that its content “may not always be error-free.” According to Reuters, the company also stated that the inaccurate information in the fake shooting story originated from an article on a site called FindPlace.xyz, rather than being a hallucination by AI.

The article was written by “Amelia Washington,” who appears to have no public profile outside of FindPlace.xyz. Elements of her profile photo were also used on unrelated sites, suggesting the use of a stock photo, Reuters discovered.