Meta Pauses Muse Chat Sharing After Researchers Find AI Conversations Publicly Accessible

Meta Halts Muse Link Sharing

Meta has paused a feature in Muse that allows users to share conversations through web links, after security researchers discovered the share URLs were publicly accessible on the web. Muse is Meta's general-purpose AI assistant, launched last month. The finding, first reported by Ars Technica, stems from a report published this week by the security firm Claroc.

"People have the ability to share specific Muse artifacts, which creates a page that can be accessed by those with the link," Muse spokesperson Tracy Clayton said in an emailed statement to The Verge. "However, we've paused this functionality while we review this report and consider changes to how these links can be discovered."

Claroc said it could no longer open a sampling of the links it had previously accessed after Meta's change.

Thousands of Links Found in Public DNS Records

According to Claroc, researchers located more than 6,800 Muse share links by searching public DNS records, which serve as a directory for web traffic. The firm said it was able to view hundreds of Muse conversations through this method.

Claroc explained that the way Meta designed the links made them easy to find. Each share's web address is built from the share's title, a short phrase describing the conversation, and then a random ID. The hostname often revealed what a page contained before it was opened, and because the link carries that same title and ID, a hostname pulled from DNS can be used to rebuild the full link.

Sensitive Conversations Exposed

Among the conversations Claroc accessed were a marketing company's revenue review, a small business owner's loan application, a bank fraud alert with partial account details, a workplace's full-year staff calendar, notes for a job candidate's call with an AI lab, and other internal company documents.

Meta CEO Mark Zuckerberg said last month that Muse was "built from the ground up for privacy and security." The incident is not the first time an AI sharing feature has made chats more public than intended. Conversations shared through OpenAI's ChatGPT, Anthropic's Claude, and xAI's Grok have also wound up on Google Search.

Sources: The Verge · Ars Technica · Engadget