Meta Pulls AI-Generated News Feed After Questions Raised

Meta has pulled an experimental feature from its standalone AI app that generated clickbait-style articles using artificial intelligence, complete with fabricated stories and distorted images of public figures.

The feature, called "For You," appeared in the Meta AI app and presented users with a feed of suggested article prompts. When tapped, these prompts generated entire stories on topics ranging from British tea etiquette to luxury watches, depending on what the algorithm determined about the user's interests.

Fabricated content and royal clones

The AI-generated articles contained little substantive information and lacked any sourcing. One story about a royal butler settling a tea debate appeared to reference a 2018 BBC comedy series, while another first-person narrative about a fake Rolex experiment seemed entirely fabricated.

The accompanying images proved particularly problematic. One article about the British royal family featured two Queen Elizabeth IIs, despite her death several years prior. The image also included distorted approximations of other royals, with faces and bodies at unnatural angles. Another image showed an older couple in a GIF making physically impossible arm movements.

None of the content carried labels indicating it was AI-generated, despite Meta's stated policy of wanting people to know when they see AI-made posts.

Why it matters

This incident reveals the tension between Meta's AI ambitions and content integrity standards. The company is pushing AI features across its platforms while simultaneously struggling to implement guardrails that prevent the spread of fabricated information and manipulated images of real people. The feature's quiet removal after press inquiry suggests Meta recognized the reputational risk of AI-generated clickbait that could easily be mistaken for legitimate content.

Rapid reversal after inquiry

When questioned about the feature, Meta initially provided vague statements about testing "a daily feed that proactively shares tips, content, and recommendations." The company declined to answer specific questions about whether it considered the output news or fiction, what safeguards existed, or whether images of real people complied with its own AI policies.

Hours later, Meta spokesperson Tracy Clayton announced the feature would be "deprecated" with no plans to move forward. The company characterized it as "a test for a limited number of users," though multiple reporters at The Verge had access to the same functionality.

The Meta AI app first launched in April 2025 with a public "Discover" feed showing AI-generated images and conversations from other users. That feature has since disappeared, replaced by a standard chatbot interface.

The Verge first reported these details after testing the feature and discovering the fabricated content and problematic images.