Meta has confirmed its decision to discontinue a recently implemented feature within its Meta AI application, following inquiries from The Verge regarding its functionality and content.
While Facebook platforms have long been associated with feeds populated by sensationalized content, Meta recently ventured into generating its own clickbait-style articles using artificial intelligence.
The standalone Meta AI app introduced a "For You" section, designed to present users with a curated list of stories reminiscent of clickbait. However, the topics, accompanying images, and textual content of these articles were entirely AI-generated, exhibiting the variable quality often associated with AI-created material.
Initially launched in April 2025, the Meta AI app first focused on a public "Discover" feed, showcasing AI-generated images and conversations from other users, some of whom appeared unaware of the public nature of their content. This "Discover" feed has since been removed. The app now features a conventional chatbot interface alongside the "For You" page, which has been operational for several months, presenting a stream of suggested article prompts that, when selected, generate full "stories."
For a reporter based in London, the AI-generated prompts were distinctly British, covering subjects such as tea customs, etiquette, pubs, the Royal Family, football, and the cultural nuances of queuing. Examples of suggested stories included "A royal butler finally settled the milk first debate" (apparently, milk goes in last), "The psychology of joining a queue without knowing why," "The anatomy of the devastating British tut," and "Inside the extreme sport of visiting every UK pub." Some prompts were even more abstract, such as "When a bit of a pickle means total disaster."
Conversely, a colleague of the reporter appeared to have been categorized by the algorithm as a luxury watch enthusiast. Their feed displayed suggested articles titled "My fake Rolex experiment" and "The brutal math behind the Rolex waitlist illusion."
The textual content produced by the AI was largely superficial, resembling verbose filler that merely reiterated the initial prompt without offering substantial depth or original information. Crucially, no sources were provided for any of the generated content.
An investigation into the origins of these "stories" revealed inconsistencies. The story about the royal butler and tea, for instance, seemed to draw inspiration from "Miss Holland," a 2018 BBC Three comedy series. In contrast, the "Rolex experiment" story appeared to be a complete fabrication, generated as a first-person narrative within the chat interface without any attribution, following the typical processing delay of a chatbot. Other articles relied on vague allusions to unnamed experts or fictional research.
Repeatedly tapping the same article card yielded stories that remained consistent with the prompt's general theme but presented slightly varied versions. However, typing the exact same headline into a separate chat session produced an entirely different response. A significant insight came from the chat history, which exposed the hidden, internal prompts used to trigger article generation. One such prompt began: "You are a helpful conversational assistant. The user is responding to a proactive feed card that was shown to them. The card context below provides background on what prompted the user’s message," followed by what appeared to be references to internal instructions, data, and metadata.
The articles were accompanied by images. While many were innocuous, depicting generic cartoonish figures, landscapes, or food, some controversially featured real individuals, including public figures, replete with inaccuracies. For example, an image illustrating "Who really pays for the royal family in 2026?" depicted two versions of Queen Elizabeth II, despite her singular existence and passing several years prior.
Surrounding these duplicate Queen Elizabeth II figures were approximations of other royals: a vaguely Princess Kate-like face, a peculiar rendering of Prince William, and a King Charles figure with an exaggerated resemblance to his late father. Other images displayed characteristic AI distortions, such as anatomically impossible hands and figures positioned at unnatural angles. One image was even a GIF of an elderly couple performing arm movements that defy human physiology.
It remained unclear whether the app's ability to generate AI images of real people aligned with Meta's own, somewhat ambiguous, guidelines. Despite Meta's previous statements about wanting "people to know when they see posts that have been made with AI" and its practice of automatically adding labels to certain user-generated AI content, there was no visible indication or label within the "For You" feed or the articles themselves to denote that the material was AI-generated.
Meta declined to provide answers to numerous questions concerning the feature's purpose, whether its output was classified as news or fiction, the safeguards in place, or if images of real individuals and public figures complied with the company's AI-content policies.
"We’re testing a daily feed that proactively shares tips, content, and recommendations tailored to your interests," stated Meta spokesperson Tracy Clayton in an initial brief statement. "The goal is to suggest what’s most relevant to you – such as fitness advice, meal plans, or other insights – before you even have to ask."
Ms. Clayton subsequently issued an almost identical "updated" statement, notably omitting the word "proactively."
A third statement from Ms. Clayton later that day confirmed: "This was a test for a limited number of users and it will be deprecated. Meta has no plans to move forward with this feature."
These developments raise further questions: How was this test considered "limited" when at least three colleagues of the reporter at The Verge, in addition to the reporter themselves, had access to the same AI clickbait feature? What precisely did "proactively" imply? And fundamentally, was there any discernible user demand for such a feature in the first place?
The Editorial Staff at AIChief is a team of professional content writers with extensive experience in AI and marketing. Founded in 2025, AIChief has quickly grown into the largest free AI resource hub in the industry.