Researchers Find Meta Muse AI Shares Sensitive User Data
· Technology · Ars Technica, TechCrunch
Security researchers at Claroc have discovered that Meta's new AI agent, Muse, exposes sensitive user information through shared conversation links. The platform generates URLs to share content, but a design flaw makes these links easily discoverable via passive DNS searches. Researchers successfully accessed 6,800 links, uncovering private data including loan applications, internal revenue reviews, and staff vacation calendars. Meta has not yet publicly addressed the specific findings regarding the security of these shared links.
Why it matters
The vulnerability exposes personal and business secrets of users who assumed their shared AI conversations were private and secure.
Read the original report — Ars Technica
Join us on Telegram
Breaking news the moment it lands. At 10,000 members we ship the Android app.