NewsBreak, the most downloaded news app in the United States, has been found to have published dozens of AI-generated fake stories over the past three years, Reuters reports.
Report: NewsBreak Is Using AI to Fabricate News Stories
According to Reuters, NewsBreak has published a fabricated article titled "Christmas Day Tragedy Strikes Bridgeton, New Jersey Amid Rising Gun Violence in Small Towns."
The article falsely reported a shooting incident in Bridgeton, New Jersey, which the local police department promptly debunked.
In a Facebook post, the Bridgeton police department dismissed the NewsBreak article as "entirely false," noting that it had been generated using AI technology. They criticized the app for publishing AI-generated fiction without proper verification.
Following the statement, NewsBreak removed the fabricated article, which it said was from a different source.
More Instances of Fake AI-Generated News Stories
A deeper investigation revealed that this incident was not isolated. Reuters also reported that since 2021, NewsBreak has published at least 40 false stories, many of which were generated using AI. These stories have had real-world consequences, affecting local communities.
For example, the Colorado-based charity Food to Power reported incorrect food distribution times on three occasions, causing confusion and frustration among people relying on their services.
Similarly, Harvest912, a charity in Pennsylvania, had to refute false reports of a 24-hour foot-care clinic for people experiencing homelessness. In a January email, Harvest912 stated, "You are doing HARM by publishing this misinformation - homeless people will walk to these venues to attend a clinic that is not happening."
Why Did This Happen?
NewsBreak has over 50 million monthly users and publishes licensed content from major media outlets like Reuters, Fox, and CNN, as well as content obtained through web scraping and AI.
The app aims to fill the void left by shuttered local news outlets, billing itself as "the go-to source for all things local." However, the extensive use of AI tools has led to significant errors.
However, this approach has raised significant concerns, with internal documents and emails reviewed by Reuters revealing alarm over the app's extensive use of AI. In a May 2022 memo to NewsBreak's CEO, Jeff Zheng, former Wall Street Journal executive editor Norm Pearlstine cautioned that the AI-generated content could "destroy the NewsBreak brand."
NewsBreak Responds: What's Next?
NewsBreak responded to these findings by removing the inaccurate articles and adding a disclaimer to its homepage in March, warning that its content "may not always be error-free."
The company also acknowledged using AI tools in a "limited experiment" in three US counties, which produced ten articles. NewsBreak claimed these efforts were disbanded and denied using content behind paywalls, stating they used publicly visible snippets.
NewsBreak is not alone in facing scrutiny over AI-generated content. Similar issues have surfaced in other outlets, such as CNET, which was found to publish inaccurate AI-powered finance explainer reports.
Join the Conversation