MENLO PARK, CALIFORNIA / RankWire.AI / – Following the discovery of over 6,800 share links exposed through public DNS records, Meta has suspended the chat sharing feature within its Muse AI assistant. Security researchers revealed that these links, which could potentially reveal sensitive personal and business data, were accessible to anyone with internet access. On Tuesday, October 20, Meta announced it had paused the feature while evaluating the findings and considering modifications to how shared links are discovered.
The exposure was uncovered by security firm Claroc, which identified more than 6,800 Muse share links in passive DNS data. These records keep track of Domain Name System lookups, making website addresses retrievable by anyone examining the data. Claroc indicated it reported the issue to Meta via its bug bounty program on October 8 and published its research on October 18. The figure of 6,800 pertains to the number of discovered addresses, not necessarily containing sensitive chat contents.
Researchers at Claroc examined several hundred of these conversations, with nearly all accessible without requiring login credentials. Among the exposed data were a loan application for a small business owner that included part of an account number, as well as a note referencing a bank fraud alert. Also visible were a revenue review prepared by a marketing firm for its CEO and a report from a real-estate investor on mortgages and returns. Confidential interview prep notes were among the accessible content as well.
Chat titles revealed through hostname patterns
The vulnerability stemmed from how Muse’s sharing links were constructed. Each URL contained the conversation title alongside a random identifier within the hostname. While the identifier could not be easily guessed, visiting these links triggered a DNS lookup, which passive DNS services could log. As a result, individuals who never received the link could find it stored in these databases, enabling them to access the conversation without signing in.
Despite Muse’s configuration instructing search engines and AI crawlers to ignore its share pages, researchers did not need a search engine to locate them. Over 75 saved scans of Muse share pages were stored on urlscan.io. Claroc reported that there was no evidence of data misuse. The firm did not contact affected users and was unable to determine whether others had discovered the links. Meta’s official statement did not specify if users would be notified about the exposure.
Pause implemented as Meta reviews security findings
On October 20, when Claroc re-verified 40 of the previously accessible share links, none of them loaded, and their addresses returned errors. The conversation titles remain visible in passive DNS history records. Muse launched in the United States on September 8, targeting adults aged 18 and over. The platform can connect with users’ email, calendar, payment, and health applications. In April, Meta introduced its Muse Spark model, developed by its superintelligence research division.
This incident highlights ongoing privacy concerns surrounding AI tools and the potential risks of sharing chats containing confidential data. In 2025, shared ChatGPT conversations appeared in Google search results, prompting OpenAI to remove the indexing feature that allowed this. Additionally, in August, Meta was ordered to fund teen mental health initiatives in New Mexico. Meta has yet to specify when Muse’s sharing feature will be reinstated or how the link security measures will be adjusted.