MENLO PARK, CALIFORNIA / RankWire.AI / – Meta has suspended conversation sharing in its Muse artificial intelligence assistant after cybersecurity researchers found that thousands of shared chat links could be discovered through public internet records. The links potentially exposed sensitive personal and corporate information. Meta told Ars Technica on Tuesday, October 20, that it had paused the feature while it reviews the findings. The company said it is also considering changes to how shared links can be discovered.
Security firm Claroc identified more than 6,800 shared Muse links through passive DNS records. These databases log domain-name lookups and can make web addresses searchable. The figure reflects discovered link addresses rather than the number of sensitive conversations. Claroc published its research on October 18. The firm said it first reported the issue to Meta’s bug bounty program on October 8, with reproduction steps and example shares.
Researchers opened several hundred conversations, and almost every live share loaded without a Meta account. Some contained sensitive information, including a marketing company’s internal revenue review and a small-business loan application. Others showed a bank fraud alert with a transaction amount and partial account details. Researchers also found confidential executive-search interview notes and a workplace vacation calendar listing dozens of employees. Many other shared chats contained routine material.
Link design left chats discoverable
The flaw stemmed from the way Muse built its share addresses. Each link combined the conversation’s title with a randomly generated identifier inside the web address’s hostname. The identifier was long enough to resist guessing. However, opening a link triggered a DNS lookup, which passive DNS services could record and make searchable. Others could then find the addresses without receiving them directly and view the chats without signing in.
Muse asked search engines and AI crawlers not to index its share pages, but that did not stop access through DNS records. The security website urlscan.io also held more than 75 saved scans of Muse share pages. Claroc said it found no evidence of misuse. The firm did not contact affected users and could not determine whether others had found the links. Meta’s statement did not say whether affected users would be notified.
Shared links now disabled after review
Claroc rechecked 40 shares on October 20 that had previously opened without an account. None of them loaded, and the addresses serving the content returned errors. Conversation titles, however, remain in passive DNS history. Meta launched Muse in the United States on September 8 for users aged 18 and older. The assistant can connect to users’ email, calendar, payment and health apps. Meta unveiled its Muse Spark model from its superintelligence lab in April.
The incident highlights growing concerns about data privacy in AI applications and the risks of sharing conversations that contain confidential information. Similar problems surfaced in 2025, when shared ChatGPT conversations appeared in Google search results. Meta has also faced legal pressure over user safety, including an order to pay $567 million following a New Mexico court ruling in August. The company has not said when Muse sharing will return.