Meta halts Muse sharing after security researchers discover publicly accessible links

Meta has temporarily disabled a feature in its Muse AI assistant that allows users to share conversations, following findings by security researchers. According to a report by Ars Technica, researchers from Claroc identified over 6,800 Muse links accessible via public DNS records. The discovery raised concerns about the potential exposure of user data. Tracy Clayton, a Muse spokesperson, stated in an emailed statement to The Verge that the feature was paused while the company reviews the report and considers changes to link management. The incident highlights ongoing privacy concerns with AI assistants, as a single query can reveal thousands of conversations. The company has not yet announced a timeline for resuming the feature or implementing new safeguards. The issue has sparked discussions about the security implications of AI-generated content being shared online. While Meta has not commented on the broader impact of the findings, the incident underscores the need for stronger data protection measures in AI platforms. Researchers emphasize that the accessibility of these links poses a significant risk to user privacy, especially for those who may inadvertently share sensitive information. The situation remains under review, with no immediate plans to restore the sharing functionality.