Meta says it will remove AI-generated story feed in Meta AI app

In short: Meta tested a “For You” feed in its Meta AI app that generated clickbait-style stories with AI, and now says it will remove it.

What happened

The Verge reports that the standalone Meta AI app includes a “For You” section that suggests headlines. When a user taps one, the app generates a full story and an image using AI (software that makes text and pictures by predicting what should come next, like a very advanced autocomplete).

The suggested headlines were tailored to the person using the app. A Verge reporter in London saw very British prompts about tea, queuing, pubs, and the royal family. Another colleague saw prompts focused on luxury watches, including stories like “My fake Rolex experiment.”

The Verge says the writing often looked like filler, with no clear sources or bylines. Some stories appeared to borrow from existing media, while others seemed made up. The attached images also included obvious mistakes, like a picture of the royal family showing two Queen Elizabeth IIs, even though there was only one and she died in 2022.

The report also says there was no obvious label in the feed or articles to tell readers the content was AI-made. Meta declined to answer many questions about safeguards and whether the content should be treated as news or fiction.

Meta spokesperson Tracy Clayton said the company was testing a daily feed of tips and recommendations. Later, Clayton said it was a limited test and that Meta “has no plans to move forward with this feature” and it will be deprecated, meaning turned off.

Why it matters

AI-written stories can look confident even when they are wrong, and they can be shared like real articles. Clear labels and sourcing help people know what they are reading, the same way a food label tells you what is inside.

Source: The Verge AI