Meta Pulls AI‑Generated Clickbait Feed After Test
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Meta added a “For You” section to its standalone AI app that automatically generates clickbait‑style story prompts, complete with AI‑written text and images.
- The feed tailors prompts to individual users, showing a London reporter British‑themed topics (tea, pubs, royals) and a colleague luxury‑watch‑centered headlines.
- The generated stories contain filler prose, no citations, and sometimes reference fictional or outdated sources such as a 2018 BBC Three comedy or a fabricated Rolex experiment.
- AI‑generated images in the feed depict real public figures—including Queen Elizabeth II and other royals—with noticeable errors and unrealistic anatomy.
- Meta’s policy states AI‑created posts should be labeled, yet the “For You” feed displayed no such disclosures on the stories or images.
- Tracy Clayton told The Verge the feature was a limited‑user test that will be deprecated, and that Meta has no plans to continue it.
- Meta declined to answer follow‑up questions about the feature’s purpose, classification as news or fiction, safeguards, and compliance with its AI‑content rules.
Why it matters: Users of Meta’s AI app were exposed to AI‑generated stories and images presented without any AI disclosure, potentially misleading readers and undermining Meta’s own labeling policy; the subsequent shutdown highlights a gap between the company’s stated safeguards and the reality of its experimental features.
