MENLO PARK, CALIFORNIA / RankWire.AI / – Following the discovery that thousands of Muse AI chat links could be accessed via public DNS records, Meta has temporarily suspended its conversation-sharing feature in the Muse assistant. The security concern raised the possibility of exposing sensitive personal and corporate information. On Tuesday, October 20, Meta announced it had halted the feature while evaluating the findings and considering modifications to how shared links are identified.
The investigation was conducted by security firm Claroc, which identified more than 6,800 Muse share links in passive DNS records. These records log Domain Name System lookups, potentially making website addresses publicly accessible. Claroc informed Meta through its bug bounty program on October 8, and published its research on October 18. The 6,800 figure pertains to the number of addresses found, not the total number of sensitive conversations accessed.
A review of several hundred shared conversations revealed that nearly all open links could be accessed without authentication. Among the data uncovered were a small-business owner’s loan application, which included part of an account number and a note about a bank fraud alert. Other examples included a marketing firm’s revenue review prepared for its chief executive and a real-estate investor’s report detailing mortgage information and returns. Confidential interview notes also appeared accessible.
Chat titles exposed through hostname leaks
The issue stemmed from the way Muse’s sharing URLs were designed. Each link embedded the conversation’s title and a unique, randomly generated identifier within the hostname. While the identifier was not straightforward to guess, visiting the link triggered a DNS lookup that passive DNS services could log. As a result, individuals who had never received the link could potentially find it via these databases. Importantly, the shared conversations could be viewed without any login requirement.
Although Muse’s settings instructed search engines and AI crawlers to ignore its shared pages, researchers did not need to rely on search engines to locate them. Over 75 saved scans of Muse share pages were stored on urlscan.io. Claroc clarified that it found no evidence of data misuse. They did not notify affected users directly and stated they could not determine if others had already discovered these links. Meta’s official response did not clarify whether users would be informed about the exposure.
Sharing feature suspended as Meta investigates
On October 20, when Claroc re-examined 40 of the previously accessible shares, none opened and all returned errors. The conversation titles remain visible in passive DNS logs. Muse was launched in the United States on September 8 for users aged 18 and older. The platform can connect with users’ email, calendar, payment, and health applications. In April, Meta introduced its Muse Spark model, created by its superintelligence lab.
This incident highlights ongoing privacy concerns related to AI tools and the risks associated with sharing chats that may contain confidential data. For example, in 2025, shared ChatGPT conversations appeared in Google search results, prompting OpenAI to remove the feature that made them indexable. Additionally, in August, Meta was ordered to fund teen mental health programs in New Mexico. Meta has not yet announced when Muse’s sharing function will be restored or how the links will be handled in the future.