Meta AI is not only a chatbot. It also has a social layer where some conversations can be surfaced publicly, and that design is drawing scrutiny because the posts can include deeply personal details tied to visible user identities.
The issue is not that every chat automatically appears in public. According to Meta, users have to choose to share a conversation. The concern is that the public Discover feed still contains examples that look less like social posts and more like private requests for help.
How Meta AI became a public feed
Meta AI launched in April as a chatbot platform with a Discover tab. In the app, that tab shows a timeline of other people’s interactions with the chatbot. On the Meta AI website, users can also see a broad collage of shared prompts and answers.
Some of what appears there is ordinary chatbot material, such as trip planning or recipe advice. But other posts include locations, telephone numbers, and sensitive personal information. The source article says these can appear alongside user names and profile photos.
That combination matters. A question that might feel casual inside a chat window can become much more revealing when it is attached to a real profile. Even if the post was shared through a deliberate process, readers may not fully understand how much context they are making visible.
The private details showing up in public
The examples reported from the public feed cover several sensitive categories. One user asked for help with a renter’s tenancy termination format. Another requested an academic warning notice that included personal details and a school’s name.
Other examples moved into legal and financial territory. One person asked about their sister’s liability in potential corporate tax fraud in a specific city, using an account connected to an Instagram profile that displayed a first and last name. Another asked for help drafting a character statement to a court while including personally identifiable information about the alleged criminal and the user.
Medical details also appeared. The source describes users asking about bowel movement problems, hives, and a rash on their inner thighs. One user discussed neck surgery and included their age and occupation.
Many, but not all, of the accounts appeared to connect to public Instagram profiles. That makes the privacy issue larger than a single prompt. A shared chatbot exchange can link a question, an identity, and a public social profile into one discoverable package.
What Meta says about sharing
Meta spokesperson Daniel Roberts told WIRED in an emailed statement that chats with Meta AI are private unless users go through a multistep process to share them on the Discover feed. The company did not respond to questions about what mitigations are in place for sharing personally identifiable information on the Meta AI platform.
Meta has also said in a company blog post that “nothing is shared to your feed unless you choose to post it.” The same blog post says users can tell the AI to “remember certain things about you” and that Meta AI can provide more relevant answers by drawing on information users have already chosen to share on Meta products, including profile information and content they like or engage with.
Those statements point to the central tension. Meta frames sharing as a user choice. Critics are focused on whether users understand the setting, the audience, and the possible consequences before they press the button.
Why privacy experts are worried
Calli Schroeder, senior counsel for the Electronic Privacy Information Center, told WIRED that she has seen people “sharing medical information, mental health information, home addresses, even things directly related to pending court cases.”
“All of that's incredibly concerning, both because I think it points to how people are misunderstanding what these chatbots do or what they're for and also misunderstanding how privacy works with these structures,” Schroeder says.
Her concern is broader than the Discover feed alone. Chatbots can feel like private assistants, especially when users ask for help with health, relationships, legal wording, school notices, or personal disputes. But the source article makes clear that at least some users are making these exchanges public, whether intentionally, mistakenly, or as part of trolling after news coverage began.
Schroeder also warned against treating AI chats as confidential spaces. “People really don't understand that nothing you put into an AI is confidential,” she says. “None of us really know how all of this information is being used. The only thing we know for sure is that it is not staying between you and the app. It is going to other people, at the very least to Meta.”
The bigger AI rollout problem
The privacy debate comes as Meta continues to push deeper into artificial intelligence. CEO Mark Zuckerberg recently announced that Meta’s AI assistant has 1 billion users across the company’s platforms. It has also been reported that Meta is creating a new AI lab led by Scale AI cofounder Alexandr Wang, dedicated to building superintelligence.
That scale raises the stakes for product design. A feature that exposes a small number of sensitive posts can still matter when the service is connected to a large ecosystem of users and profiles. The issue is not only whether a setting exists, but whether people understand it in the moment they use it.
One public query captured the problem directly. A user asked Meta AI on Thursday, “Is Meta aware of how much sensitive information its users are mistakenly making publicly available,” and that question itself appeared in the public feed.
The chatbot’s answer was blunt about the challenge: “Some users might unintentionally share sensitive info due to misunderstandings about platform defaults or changes in settings over time,” it responded. “Meta provides tools and resources to help users manage their privacy, but it’s an ongoing challenge.”
For users, the practical lesson is simple: a chatbot prompt can contain more personal data than expected. In Meta AI, a shared chat may not just reveal the question. It may also reveal who asked it.