MENLO PARK, CALIFORNIA / RankWire.AI / – Security researchers uncovered more than 6,800 Muse chat links stored in passive DNS records, prompting Meta to temporarily suspend the chat sharing feature in its Muse AI assistant. The exposure risked revealing sensitive personal and business data. On Tuesday, October 20, Meta announced it had paused the function while reviewing the findings. The company is considering adjustments to how shared links can be accessed.
The discovery was made by security firm Claroc, which identified the multitude of Muse share links within passive DNS data. These records contain Domain Name System queries, which can make website addresses available to anyone inspecting the data. Claroc informed Meta through its bug bounty program on October 8 and made its research public on October 18. The 6,800 figure represents the number of found addresses, not individual conversations or sensitive content.
A few hundred of these conversations were examined by Claroc’s researchers, with nearly all of the shared links opening without requiring login credentials. Among the exposed data were a small-business owner’s loan application containing part of an account number and a note concerning a bank fraud alert. Additional shared content included a marketing firm’s revenue review prepared for its CEO and a real-estate investor’s report on mortgages and returns. Confidential interview prep notes were also visible in some instances.
Chat titles exposed through hostname leaks
The core issue stemmed from Muse’s sharing address structure. Each link incorporated the conversation title and a random identifier in the hostname. While the identifier could not be easily guessed, visiting such links triggered DNS lookups, which passive DNS services logged. This meant anyone who knew or guessed the link could retrieve it from these logs, even if they had not received the link directly. The shared chats could be accessed without requiring user login.
Muse’s configuration instructed search engines and AI crawlers to ignore its share pages, yet researchers did not rely on search engines to locate these links. Over 75 saved scans of Muse share pages were stored on urlscan.io. Claroc stated it found no evidence that anyone misused the data. The firm did not reach out to affected users and admitted it could not determine whether others had discovered the links. Meta’s official comment did not clarify if or when users would be notified about the incident.
Suspension implemented as Meta reviews findings
When Claroc re-examined 40 of the previously accessible shares on October 20, none of the links worked and all returned errors. Despite this, the conversation titles remain visible within passive DNS logs. Muse launched in the United States on September 8 for users aged 18 and older. It offers integration with email, calendar, payment, and health applications. Earlier, Meta introduced its Muse Spark model, developed by its superintelligence lab, in April.
This incident highlights ongoing privacy concerns surrounding AI tools and the dangers of sharing chats containing confidential information. Notably, in 2025, shared ChatGPT conversations appeared in Google search results, leading OpenAI to remove the feature that allowed indexing. Additionally, Meta was ordered to fund teen mental health programs in New Mexico in August. Meta has not yet announced when Muse’s sharing feature will be restored or how its link-sharing mechanisms might be altered.