Meta Pauses Muse Sharing After Reports It Exposed Users' Private Details
Privacy·October 21, 2026

Meta has switched off the sharing feature in Muse while its privacy team looks into a problem that could expose people's personal and business information. The concern is straightforward: when a user shares a conversation, the material in that conversation can travel further than the user expected, and in this case it appears to have included details the user never meant to make visible.
The feature was built to make it easy to pass a chat along to someone else, whether that is a colleague reviewing a draft or a friend looking at an answer. That convenience is also where the risk sits. People often type sensitive material into an AI assistant without thinking of it as a document, including account details, internal plans, client names or health information. A sharing button that carries the full context of a chat can turn a private exchange into something much more public than the person who wrote it had in mind.
Meta's decision to suspend the feature rather than patch it quietly is a meaningful signal. Pausing a function while an investigation is still open is a way of stopping further exposure while the company works out what went wrong and how far it reached. The company has not, based on the material available here, said how many shared conversations were affected or how long the problem existed. Those are the questions that will determine how serious the episode is for users and for Meta's regulators.
For anyone who has used sharing features in AI tools, the episode is a reminder to check what a shared link actually contains before sending it on. Review the conversation for anything personal or confidential, and treat any link you have already shared as something that may have carried more than you intended. For product teams across the industry, the lesson is that sharing tools need the same scrutiny as any other data export. A feature that makes a chat portable should make its contents clear and its boundaries explicit before it ships.
Reporting based on an external source.