Home / Technology / Meta's Muse AI Leaks Conversations Via Public Links
Meta's Muse AI Leaks Conversations Via Public Links
20 Oct
Summary
- Meta paused its Muse AI feature due to publicly available conversation URLs.
- Researchers found over 6,800 Muse links in public DNS records.
- Similar privacy issues have affected other AI chatbots like ChatGPT.
Meta has temporarily halted a feature within its Muse AI assistant that allowed users to share conversations. This decision follows a report by security firm Claroc, which found that URLs for these shared conversations were publicly accessible online. Researchers identified more than 6,800 Muse links through public DNS records.
The Muse feature enabled users to generate a shareable link for specific conversations. However, Claroc detailed how the construction of these web addresses, combining titles, conversation descriptions, and random IDs, made them discoverable via DNS lookups. This allowed access to varied sensitive discussions, including financial reviews and personal applications.
This privacy lapse is not isolated; similar issues have previously affected AI tools like ChatGPT, Claude, and Grok, where conversations inadvertently became visible through search engines. Meta is now reviewing the report and considering modifications to prevent future link discovery.
