How private is your AI search history? If you’re using Meta AI, the answer might be not as private as you think.
Meta’s newly launched artificial intelligence tool is raising eyebrows after it emerged that some user prompts and responses are being displayed publicly potentially without users fully realising it. From test questions to deeply personal topics and even requests for inappropriate content, people may be unintentionally revealing far more than they intended, with some posts traceable directly to their social media profiles.
AI Searches Go Public Sometimes with Real Names Attached
Meta AI, which integrates across Facebook, Instagram, and WhatsApp, is also available as a standalone product with a “Discover” feed where users can explore others’ interactions with the chatbot.
The problem? Many people appear unaware that their chats are being posted to this feed. Despite a warning that states, “Prompts you post are public and visible to everyone… Avoid sharing personal or sensitive information.” Users are continuing to share queries that many would expect to remain private.
The BBC found multiple instances of users uploading photos of school and university tests and asking Meta AI for answers. One chat, labelled “Generative AI tackles math problems with ease,” showed how students may be using the tool to skirt academic integrity rules.
More concerning still were conversations involving deeply personal topics, including gender identity exploration. In one example, a user publicly shared a conversation with the AI, questioning whether they should transition, raising fears that private struggles are being exposed online without full consent.
From Inappropriate Content to Identity Exposure
In some cases, the nature of the content has sparked further alarm. Public prompts included searches for images of scantily clad women and anthropomorphic animal characters. One request, which could be traced directly to a person’s Instagram account through their profile picture and username, asked Meta AI to generate an image of an animated character lying outdoors in only underwear.
Such visibility poses not only reputational risks but also the potential for harassment or misuse of the shared data.
Rachel Tobac, CEO of cybersecurity firm Social Proof Security, called the situation “a huge user experience and security problem.” In a post on X (formerly Twitter), Tobac warned that “if a user’s expectations about how a tool functions don’t match reality, you’ve got yourself a huge user experience and security problem.”
She noted that AI interactions are generally assumed to be private and not something that would appear on a public feed alongside social media content. “Because of this, users are inadvertently posting sensitive info to a public feed with their identity linked,” she said.
Meta Says Users Are “In Control,” But Are They?
In its April press release introducing Meta AI, the company stated, “You’re in control: nothing is shared to your feed unless you choose to post it.” It also described the Discover feed as “a place to share and explore how others are using AI.”
Meta says that chats are private by default and users can choose to make a post public or delete it later. There’s also the option to make searches private via account settings.
However, the way the interface is designed and the placement of the disclosure message may not make it obvious to all users when they are making something public.
With Meta AI still relatively new, especially outside the US, where it’s currently available via browsers, these early reports raise important questions about how clearly companies communicate privacy features to users, particularly when AI is involved.
A Privacy Reckoning for AI-Enhanced Social Platforms
As artificial intelligence continues to be woven into social platforms, the boundary between private queries and public posts grows increasingly blurry.
While Meta insists users have control, the current backlash highlights a gap between user expectation and platform design, a disconnect that could have lasting consequences for those unwittingly exposing their personal lives to the internet.
At a time when AI is touted as a transformative force, ensuring transparency and user comprehension around data sharing must be more than just an afterthought; it’s an ethical imperative.
