MENLO PARK, CALIFORNIA / RankWire.AI / – Meta has put a stop to sharing conversations through its Muse AI assistant after cybersecurity experts revealed that thousands of shared chat links could be retrieved from public internet records. These links potentially revealed sensitive personal and corporate data. On Tuesday, October 20, Meta informed Ars Technica that the feature was temporarily disabled while the company examines the discovered vulnerabilities. Additionally, Meta is contemplating modifications to how shared links can be accessed or discovered.
Security firm Claroc identified over 6,800 shared Muse links by analyzing passive DNS records, which log domain name queries and can be used to search for web addresses. This number reflects the discovered link addresses rather than the total number of sensitive conversations. Claroc published its findings on October 18 and stated that it first reported the issue to Meta’s bug bounty program on October 8, including steps to reproduce the problem and examples of shared links.
The researchers accessed several hundred conversations, with nearly all active shares opening without requiring a Meta account. Some of these contained confidential information, such as an internal revenue review for a marketing firm and a small-business loan application. Other shared chats displayed a bank fraud alert, including transaction amounts and partial account details. The team also uncovered confidential interview notes for executive-search candidates and an employee vacation calendar. Many other shared conversations consisted of routine content.
Vulnerable link structure caused chat discoverability
The security flaw originated from the way Muse generated its share URLs. Each link combined the chat’s title with a long, randomly generated identifier embedded within the hostname of the web address. This identifier was sufficiently complex to prevent guessing attempts. However, when a link was opened, it triggered a DNS query, which passive DNS services could record and index, making the URLs searchable. This allowed others to find the chat addresses without direct access and view conversations without signing in.
Although Muse instructed search engines and AI crawlers not to index its share pages, this did not prevent access via DNS records. The security site urlscan.io stored more than 75 scans of Muse share pages. Claroc stated that it found no evidence of misuse but did not contact the affected users and could not determine whether anyone else discovered the links. Meta’s official statement did not specify whether users impacted by the issue would be notified.
Sharing links disabled following review and investigation
On October 20, Claroc re-verified 40 shares previously accessible without an account; none of these loaded, and the content-serving addresses returned errors. The conversation titles, however, remain stored in passive DNS logs. Meta launched Muse in the United States on September 8, targeting users aged 18 and older. The AI assistant can connect to personal email, calendars, payment, and health applications. In April, Meta introduced its Muse Spark model developed by its superintelligence research division.
This incident underscores increasing concerns over data privacy in AI tools and the risks associated with sharing confidential conversations. Similar issues arose in 2025 when shared ChatGPT discussions appeared in Google search results. Meta has also faced legal challenges related to user safety, including an order to pay $567 million following a New Mexico court ruling in August. The company has not announced when Muse sharing features will be restored.