MENLO PARK, CALIFORNIA / RankWire.AI / – Meta has temporarily halted the sharing feature within its Muse AI assistant following security research revealing that over 6,800 shared chat URLs were accessible via public internet DNS records. This vulnerability could have potentially exposed private personal and business information. On Tuesday, October 20, Meta announced it had suspended the feature while it investigates the issue further. The company is currently considering modifications to how shared links are discoverable.
The discovery was made by cybersecurity firm Claroc, which identified more than 6,800 Muse share links within passive DNS records. These records store Domain Name System lookups and can make website addresses accessible to anyone conducting searches on the stored data. Claroc stated it informed Meta through its bug bounty program on October 8, and published its findings on October 18. It is important to note that the 6,800 figure pertains to discovered addresses, not sensitive conversation content.
Researchers from Claroc examined a few hundred of these shared conversations, almost all of which were accessible without requiring login credentials. Among the exposed data were a small-business owner’s loan application that included part of an account number, as well as a note regarding a bank fraud alert. Other shared information included a marketing firm’s revenue analysis prepared for its CEO, and a real-estate investor’s report detailing mortgages and investment returns. Confidential interview preparation notes were also among the visible content.
Shared Chat Titles Exposed via Hostname Configuration
The core issue stemmed from how Muse’s sharing URLs were constructed. Each link embedded the conversation title along with a random identifier within the hostname. Although this identifier was not easy to predict, visiting a link triggered a DNS lookup that passive DNS services could log. Consequently, individuals who never received a direct link could still locate it within these databases. The shared conversations could be accessed freely, with no sign-in required.
While Muse’s privacy settings instructed search engines and AI crawlers to avoid indexing share pages, this did not prevent passive DNS records from revealing them. Over 75 saved snapshots of Muse share pages are stored on urlscan.io. Claroc clarified that there is no evidence to suggest that the data was misused by anyone. The firm did not reach out to affected users and could not determine if others had discovered the links. Meta’s official statement did not specify whether users would be notified about the exposure.
Sharing Feature Suspended During Review Process
Following a re-examination of 40 previously accessible shares on October 20, Claroc found that none of these links loaded, and their addresses returned errors. Despite this, the conversation titles remain visible in passive DNS records. Muse was launched in the United States on September 8, targeting users aged 18 and older. The app can integrate with users’ email, calendar, payment, and health apps. Additionally, Meta introduced its Muse Spark model, developed by its superintelligence research division, in April.
This incident underscores ongoing concerns regarding privacy in AI tools and the risks linked to sharing chats that may contain sensitive data. In 2025, shared conversations from ChatGPT began appearing in Google search results, prompting OpenAI to remove the feature that allowed indexing. Furthermore, in August, Meta was ordered to fund teen mental health programs in New Mexico. The company has not yet specified when Muse’s sharing capabilities will be restored or how it intends to modify link security measures.