GoKawiilMeta restores Muse AI sharing after fixing exposed artifact URLs
Meta re-enabled sharing on its Muse AI assistant after temporarily disabling it when security researchers found that shared conversation links were predictably named and discoverable via public DNS records. Meta executive David Singleton said the flaw affected URLs for shared 'artifacts' like documents and images, not chat sharing itself, and has now been fixed. Previously shared artifact links have been invalidated and given new URLs, requiring users to re-share them.
GoKawiil's interpretation of the reporting above, not reported fact.
The exposure allowed outside researchers to stumble upon sensitive user content, including a company's internal revenue review and a personal loan application, highlighting how AI tools that generate shareable links can inadvertently leak private data if URL structures are predictable. Meta's quick pause-and-fix response suggests the company is trying to balance product rollout speed with data protection, though the incident may raise broader questions about security practices in rapidly shipped AI features.
- Meta paused then restored Muse's sharing feature after a URL predictability flaw was found by security researchers.
- The bug exposed artifact links (documents, images) rather than chat conversations, per Meta's David Singleton.
- Previously shared artifact URLs are now invalid, requiring users to request and share new links.
Source: theverge.com — Jay Peters, 2026-10-22
Published there as: “Meta turns Muse sharing back on”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.