MENLO PARK, CALIFORNIA / RankWire.AI / – Meta has paused the sharing of conversations within its Muse AI assistant after cybersecurity experts discovered that thousands of shared chat links could be retrieved through publicly accessible internet records. These links potentially revealed sensitive personal and corporate data. Meta informed Ars Technica on Tuesday, October 20, that the feature is temporarily suspended while it assesses the findings. The company is also considering modifications to how shared links are made discoverable.
Security company Claroc identified over 6,800 shared Muse links via passive DNS records. Such databases log domain name lookups and enable web addresses to be searchable. This number reflects discovered link URLs rather than the total of sensitive conversations. Claroc published its findings on October 18, stating it first reported the issue to Meta’s bug bounty program on October 8, providing reproduction steps and sample shares.
Researchers accessed several hundred conversations, with almost all shared links loading without requiring a Meta account. Some contained highly sensitive details, including an internal revenue review from a marketing firm and a small-business loan application. Others displayed a bank fraud alert along with a transaction amount and partial account information. Additionally, researchers uncovered confidential interview notes for executive searches and a workplace vacation calendar listing dozens of employees. Numerous other shared chats were routine in nature.
Vulnerability caused by link design allowed chat discovery
The issue originated from the way Muse constructed its share URLs. Each link combined the chat title with a randomly generated identifier within the hostname of the web address. This identifier was lengthy enough to prevent guessing. Nevertheless, clicking a link triggered a DNS lookup, which passive DNS services could log and make searchable. As a result, others could locate these addresses without direct access and view the conversations without signing in.
Muse requested search engines and AI crawlers not to index its share pages, yet this did not prevent access via DNS records. The security site urlscan.io stored over 75 saved scans of Muse share pages. Claroc stated it found no evidence indicating misuse. The firm did not contact the affected users and could not determine whether others had discovered the links. Meta’s statement did not specify if users affected by this issue would be notified.
Sharing links are now disabled following review
On October 20, Claroc re-evaluated 40 shares that had previously loaded without a Meta account. All of these now returned errors, and the addresses serving the content could no longer be accessed. However, conversation titles remain visible in passive DNS history. Meta launched Muse in the United States on September 8, targeting users aged 18 and older. The assistant can connect to various apps such as email, calendar, payment, and health services. In April, Meta introduced its Muse Spark model developed by its superintelligence research lab.
This incident underscores the rising concerns over data privacy in AI applications and the dangers of sharing conversations that may contain confidential information. Similar issues appeared in 2025 when shared ChatGPT conversations surfaced in Google search results. Meta has also faced legal challenges regarding user safety, including an order to pay $567 million after a New Mexico court ruling in August. The company has not announced when Muse’s sharing feature will be reinstated.