Concerns are mounting over Meta’s new artificial intelligence tool after it emerged that some users may be unknowingly sharing sensitive search queries and images to a public feed, potentially linking them to their social media identities.
Meta AI, which can be accessed via Facebook, Instagram, WhatsApp, and a standalone website, allows users to interact with the tool by asking questions or generating content. But those prompts—and Meta AI’s responses—can end up in a public “Discover” feed if users choose to share them, sometimes without fully realising the implications.
A message does appear before a post is made public, warning: “Prompts you post are public and visible to everyone… Avoid sharing personal or sensitive information.” However, cybersecurity experts warn that the warning may not be clear enough for users to understand that their queries—ranging from exam cheating to highly personal topics—could become visible to anyone online.
The BBC found several examples of sensitive and potentially embarrassing queries made public, including a user exploring gender identity, requests for AI-generated semi-nude animated characters, and test questions uploaded by students seeking help. In at least one case, a user’s identity was traceable through their Meta AI post due to a visible username and profile photo linked to their Instagram account.
Cybersecurity expert Rachel Tobac, CEO of Social Proof Security, criticised Meta’s user interface, calling it “a huge user experience and security problem.” She added on social media: “If a user’s expectations about how a tool functions don’t match reality, you’ve got yourself a problem. Because of this, users are inadvertently posting sensitive info to a public feed with their identity linked.”
Meta says that chats are private by default and users are in full control of what they share. In an April press release, the company said, “Nothing is shared to your feed unless you choose to post it.” Users can delete public posts and adjust privacy settings to prevent future sharing.
Still, the ease with which some of these prompts appear in the public feed raises concerns about whether users are adequately informed. The “Discover” section is promoted as a place to explore how others are using Meta AI, but critics say many users don’t realise how visible their posts become or that their identities could be connected to them.
Meta AI is currently available via browser in the UK and through a dedicated app in the US. Users are being urged to review their privacy settings carefully before using the tool.
