Meta has added a warning label to its AI app after users were found publicly posting personal conversations without realising those interactions were being shared. The update, now visible on the app’s “post to feed” button, advises people not to include personal or sensitive information when publishing content. The change follows mounting concern over how easily private exchanges were being made publicly viewable.
Reports first emerged after users noticed that the app’s public feed, called “discover”, was showing highly personal queries submitted to Meta’s AI, ranging from health concerns to legal questions. In some cases, users even included phrases such as “please keep this private” in what were already public posts. While the app does not share conversations by default, many people appear to have been unaware that using the “share” option meant their chats would be made visible to everyone.
The volume and nature of the content being exposed drew criticism from privacy advocates, who pointed out that most AI chat apps do not display user interactions in a public-facing stream. Observers have noted that the app offers few familiar cues or design features to indicate that posts are going live. Some compared it unfavourably to Meta’s other platforms, where sharing controls are more visible and widely understood.
The new warning message now appears when users try to post a chat publicly, but based on reports, it only shows once, on the first attempt. It clearly states that prompts shared with the feed are public, may appear on other Meta platforms, and should not include private details. However, critics argue that a single disclaimer may not be enough to prevent confusion, especially for users who miss or dismiss the initial notice.
Image: Business Insider
At the same time, it appears that Meta has altered what kind of content is shown in the app’s public feed. Text-based conversations no longer seem to appear. Instead, the feed is currently displaying AI-generated images and video content. It is unclear whether this change is permanent or whether it reflects a temporary shift following the attention the app has received in recent days. Meta has not confirmed whether this is part of a wider redesign or a direct response to the criticism.
Those who have already shared personal content can still take steps to reduce their visibility. Within the app’s settings, it is possible to hide all past posts by adjusting the privacy options under the “Data & Privacy” section. Users can make every prompt private in just a few taps, though the setting may not be obvious at first glance.
The issue has sparked renewed scrutiny of how emerging AI tools are being integrated into social platforms. While many users approach chatbots expecting private assistance, the blending of conversational AI with public sharing features appears to have created misunderstandings, with very real privacy consequences.
Read next: Meta Expands WhatsApp Monetization with In-App Ads and Paid Channel Subscriptions
Reports first emerged after users noticed that the app’s public feed, called “discover”, was showing highly personal queries submitted to Meta’s AI, ranging from health concerns to legal questions. In some cases, users even included phrases such as “please keep this private” in what were already public posts. While the app does not share conversations by default, many people appear to have been unaware that using the “share” option meant their chats would be made visible to everyone.
The volume and nature of the content being exposed drew criticism from privacy advocates, who pointed out that most AI chat apps do not display user interactions in a public-facing stream. Observers have noted that the app offers few familiar cues or design features to indicate that posts are going live. Some compared it unfavourably to Meta’s other platforms, where sharing controls are more visible and widely understood.
The new warning message now appears when users try to post a chat publicly, but based on reports, it only shows once, on the first attempt. It clearly states that prompts shared with the feed are public, may appear on other Meta platforms, and should not include private details. However, critics argue that a single disclaimer may not be enough to prevent confusion, especially for users who miss or dismiss the initial notice.
Image: Business Insider
At the same time, it appears that Meta has altered what kind of content is shown in the app’s public feed. Text-based conversations no longer seem to appear. Instead, the feed is currently displaying AI-generated images and video content. It is unclear whether this change is permanent or whether it reflects a temporary shift following the attention the app has received in recent days. Meta has not confirmed whether this is part of a wider redesign or a direct response to the criticism.
Those who have already shared personal content can still take steps to reduce their visibility. Within the app’s settings, it is possible to hide all past posts by adjusting the privacy options under the “Data & Privacy” section. Users can make every prompt private in just a few taps, though the setting may not be obvious at first glance.
The issue has sparked renewed scrutiny of how emerging AI tools are being integrated into social platforms. While many users approach chatbots expecting private assistance, the blending of conversational AI with public sharing features appears to have created misunderstandings, with very real privacy consequences.
Read next: Meta Expands WhatsApp Monetization with In-App Ads and Paid Channel Subscriptions
