Meta’s Muse AI Shares Sensitive Data Through Easily Discoverable Links

Meta’s new general‑purpose AI agent, Muse, is designed to let users share searches, conversations, images, and other content via unique URLs. Researchers at security firm Claroc discovered that the URL generation method makes the links easy to enumerate. They identified 6,800 share links and opened a few hundred, finding personal and business information that users had not intended to publish. Meta’s CEO Mark Zuckerberg had previously claimed Muse was built “from the ground up for privacy and security,” but the findings suggest that the design flaw compromises that promise. The incident highlights a gap between Meta’s stated privacy goals and the practical security of its sharing features.