Meta AI Discovery Feed is reportedly filled with users' seemingly private chats

 

The Discover feed of the Meta AI app is apparently being swamped with users' private discussions and requests that have been mistakenly shared with the public. Several reports have highlighted instances of text chats and image prompts that look profoundly personal, implying that users may be mistakenly sharing them to the app's social feed. The surfacing of these public messages has sparked worry among netizens, with experts asking whether the corporation is doing enough to preserve its users' privacy.

Meta AI app's social feed reportedly contains posts on medical and legal queries
According to a TechCrunch report, posts on the Meta AI app's Discover feed include "people asking for help with tax evasion," "how to write a character reference letter for an employee facing legal troubles," and even "a bunch of red bumps on my inner thigh" questions.

The outlet, which also shared images of related public posts, is not alone. Kylie Robinson, Senior Correspondent at Wired and author of the story, claimed to have seen postings from users asking, "Do younger women like older white men?" and other questions.

Furthermore, in an interview with Wired, Calli Schroeder, Senior Counsel for the Electronic Privacy Information Center, stated that she has seen posts where individuals are "sharing medical information, mental health information, home addresses, even things directly related to pending court cases."

While Gadgets 360 team members have not seen such explicitly private postings, we have seen unusual posts in which users have shared selfies as a request to the chatbot to make a small correction. Some of these photos show minors who have altered their hair color. Furthermore, numerous X (previously known as Twitter) users have shared images of such private tweets on the app's social feed.

🚨 Think your chats are private? Meta AI’s confusing design has users accidentally sharing sensitive conversations publicly. Plus, AI chats on WhatsApp aren’t end-to-end encrypted, meaning Meta can access and use your data for training. Stay alert! 🔒🤖 #PrivacyMatters #MetaAI pic.twitter.com/KhyFILoarM

— sumant maurya (@itssumantmaurya) June 13, 2025

It cannot be certain that in all of these cases, persons were uninformed that they were making a public post. However, the nature of the inquiries and photos raises suspicions.

Currently, publicly posting a post on Meta AI requires two steps once a user has had a discussion with the chatbot. A Share icon appears at the top of the chat window, and pressing it sends the user to a new page labeled "Preview." The AI creates an editable post title, inquiry, and response, as well as a huge "Post" button at the bottom.

However, there is nothing further to inform users that pressing the Post button makes the entire chat public and available to others. While people familiar with current applications will understand what the Post button implies, individuals who are less tech-savvy may accidentally hit it and not notice they have shared a chat.

At the time of unveiling the Meta AI app, the tech titan stated, 'And as usual, you're in control: nothing is published to your feed unless you want to post it." However, the recorded cases raise concerns about whether the two-step process is sufficient to make individuals aware of the ramifications of pressing the "Post" button.