Meta has taken a step towards addressing the issue of excessive sharing on its AI app’s public feed. The company has introduced a brief disclaimer on the “post to feed” button, cautioning users to “avoid sharing personal or sensitive information.”

This change was first reported by Business Insider, which had previously described the app as “one of the most depressing places online” due to the large amount of intimate and sometimes personally-identifying information being shared publicly. Although Meta AI does not share user chat histories by default, many users were unwittingly making their interactions public by choosing to share them on the app’s built-in “discover” feed.

Recently, posts have been found where users were seeking advice on sensitive topics, such as “improving bowel movements” and inquiring about potential tax liabilities. Another user had attempted to keep their embarrassing chats private by adding a note to their public posts. These types of public interactions have been occurring since the app’s launch in April but gained renewed attention last week after social media users began sharing examples of the weird conversations visible in the app’s “discover” feed.

Privacy experts have criticized Meta, pointing out that other mainstream AI chatbots do not have a publicly visible feed. According to Rachel Tobac, a security expert who has partnered with Meta, “If a user’s expectations about how a tool functions don’t match reality, you’ve got a huge user experience and security problem.” Tobac observed that humans have a certain understanding of how AI chatbots work and do not expect their interactions to appear in a social media-style feed.

The Mozilla Foundation has also urged Meta to modify the app’s design, stating that “Meta AI’s app doesn’t make it obvious that what you share goes fully public.” The foundation noted that there is no clear iconography or familiar cues about sharing, unlike other Meta apps.

Meta has now taken action, and when users choose to share a Meta AI interaction publicly, they are prompted with a warning. The warning states, “Prompts you post are public and visible to everyone. Your prompts may be suggested by Meta on other Meta apps. Avoid sharing personal or sensitive information.” However, this warning only appears on the first share.

Additionally, the app’s public feed seems to have stopped featuring text exchanges shared by other users, only showing AI-generated images and videos. It is unclear if this is a permanent change or a result of the recent negative attention the app has received. We have reached out to Meta for more information and will update if we receive a response.

In the meantime, users who have inadvertently made public posts can remove them by navigating to their profile, then Data & Privacy -> Manage your information -> Make all public prompts visible only to you and selecting “apply to all.”