Meta AI searches made public!

How would you feel if your internet search history suddenly went public?

That’s the concern facing some users of Meta AI, as prompts submitted to the tool—along with AI-generated responses—are being displayed in a public “Discover” feed, sometimes without users fully understanding the implications.

Cybersecurity experts are warning that this poses a significant privacy and safety risk. In some cases, the public posts are easily traceable back to users' social media accounts through visible usernames and profile pictures, raising questions about informed consent and data exposure.

Some of the publicly shared prompts include requests for AI-generated scantily clad characters, help with exam questions, and deeply personal topics—such as gender identity exploration. The BBC found examples where users uploaded school and university tests, seeking answers from the chatbot. In one instance, a request linked to a person’s Instagram profile asked Meta AI to generate an image of an animated character wearing only underwear.

While Meta says that chats are private by default, users can choose to post them publicly. Before making a post public, a warning appears stating: “Prompts you post are public and visible to everyone... Avoid sharing personal or sensitive information.” Posts can be deleted after being shared.

However, critics argue that this disclaimer may not be enough. The nature of some posts suggests that users may not fully grasp that their chats are being published for public view—and possibly linked back to their social media identity.

Rachel Tobac, CEO of cybersecurity firm Social Proof Security, criticized the feature on X (formerly Twitter), saying: “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 added that most people don’t expect their interactions with an AI chatbot to appear in a feed similar to those found on social media platforms.

“Because of this, users are inadvertently posting sensitive info to a public feed with their identity linked,” Tobac warned.

Meta AI, which was launched earlier this year, is integrated into Facebook, Instagram, and WhatsApp, and is also available as a standalone web-based product. While the “Discover” feed is designed to showcase how people are using AI, it may be doing so at the expense of user privacy.

In its April press release, Meta stated: “You’re in control: nothing is shared to your feed unless you choose to post it.” However, the apparent gap between user expectations and how the tool functions is now prompting concerns.

Users can adjust their settings to keep interactions private, but the default experience—combined with the ease of accidentally sharing sensitive prompts—has raised alarms about data protection and transparency.