Muse, Meta’s new AI assistant, has a design flaw that exposes users’ private conversations when they share links. Researchers from Claroc uncovered 6,800 such links, finding sensitive information including confidential business plans, personal financial data, and vacation schedules. Meta has suspended the sharing feature while it reviews the issue.

The ease of discovery stems from the domain name and URL structure, which Claroc calls 'GhostShares.' The company’s focus on privacy and security has been called into question after this incident. Asaf Aprozper, a Claroc cofounder, reported the findings through Meta’s bug bounty program but received no response on LinkedIn.

The incident raises questions about the balance between convenience and security, especially when it comes to sharing information through AI assistants. With Muse’s rapid rise in popularity and its extensive access to user data, this flaw could have serious implications for privacy. Meta’s suspension of the feature is a temporary fix, but the underlying issues remain.