GoKawiilMeta's Muse AI shared chat links that exposed users' private data
Security researchers at Claroc found that Meta's Muse AI agent generated shareable web links for user chat logs that could be discovered and opened by outsiders. Using a lookup dataset of visited URLs, investigators retrieved over 6,800 Muse share links in a single search, revealing private material such as corporate revenue documents, loan applications, bank transaction details, income data, and a workplace vacation calendar. Meta told Ars Technica it has disabled the sharing feature.
GoKawiil's interpretation of the reporting above, not reported fact.
The exposure undercuts Mark Zuckerberg's public claim that Muse was 'built from the ground up for privacy and security,' a claim X users had already flagged with a community note. The incident suggests that AI agents designed for convenience features like shareable chat links may carry overlooked risks when URL structures can be enumerated or indexed, raising broader questions about how AI companies handle sensitive user-generated content.
- Muse's share links could be mass-discovered via a web address lookup dataset, exposing over 6,800 chats.
- Leaked data reportedly included financial records, loan applications, and personal income information.
- Meta disabled the sharing feature after researchers reported the flaw to Ars Technica.
Source: futurism.com — Joe Wilkins, 2026-10-20
Published there as: “Meta’s Muse AI Agent Caught Leaking Users’ Intimate Personal Information”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.