MENLO PARK, CALIFORNIA / RankWire.AI / – After security researchers revealed that over 6,800 shared chat links could be accessed via public internet records, Meta has temporarily halted the sharing feature in its Muse AI assistant. The exposed links potentially risked revealing private personal and business information. On Tuesday, October 20,, Meta announced the suspension of the feature while it evaluates the findings and considers adjustments to how shared links are discoverable.
The discovery was made by security company Claroc, which found more than 6,800 Muse share links in passive DNS records. These records log Domain Name System lookups, making website addresses accessible to anyone inspecting the data. Claroc reported the issue to Meta through its bug bounty program on October 8 and released its research on October 18. The figure of 6,800 refers to the number of addresses found, not to sensitive conversations.
A few hundred of these conversations were reviewed by Claroc’s team, who found that nearly all active shares could be accessed without logging in. Among the exposed data were a small-business owner’s loan application containing part of an account number, and a note about a bank fraud alert. Also uncovered were a marketing firm’s revenue report prepared for its CEO, a real-estate investor’s report on mortgages and returns, and confidential interview notes.
Hostnames revealed chat titles
The vulnerability stemmed from how Muse’s sharing URLs were constructed, with each link embedding the conversation title and a random identifier in the hostname. While the identifier was not straightforward to guess, accessing a link triggered a DNS lookup that passive DNS services could record. Consequently, anyone who didn’t receive the link directly could find it through these databases. The shared conversations could be opened without any login requirement.
Although Muse’s configuration instructed search engines and AI crawlers to ignore its shared pages, researchers did not rely on search engines to locate them. Over 75 saved scans of Muse sharing pages were stored on urlscan.io. Claroc clarified that there was no evidence of data misuse, and it did not contact users affected. It also noted that it cannot determine whether others had discovered the links. Meta’s official statement did not clarify if users would be notified about the exposure.
Sharing feature paused for review
On October 20,, Claroc rechecked 40 previously accessible shares, and none loaded, with their addresses returning errors. The conversation titles, however, still appeared in passive DNS logs. Muse was launched in the United States on September 8, for users aged 18 and over, and it integrates with email, calendar, payment, and health apps. In April, Meta introduced its Muse Spark model, developed by its superintelligence laboratory.
This incident adds to ongoing concerns about privacy issues related to AI tools and the dangers of sharing chats that contain sensitive data. In 2025, shared ChatGPT conversations appeared in Google search results, leading OpenAI to remove the option to make chats indexable. Additionally, in August, Meta was ordered to fund teen mental health programs in New Mexico. The company has not announced when Muse’s sharing function will be restored or how the link sharing process might be modified.