The incident centers on the new Muse Mac application, which boasts deep integration with Calendar, Notes, and Messages. Despite the assistant's insistence that it does not read actual message histories, it admitted to capturing notification previews through device sync. When pressed for technical specifics, the AI conceded it could not explain its own “plumbing,” leaving users to wonder about the extent of the software’s reach into their personal communications.
Meta’s Muse AI raises privacy alarms over notification access
When Jason Aten, a contributing editor at Inc Magazine, interrogated Meta’s new Muse AI about its knowledge of his private text threads, the assistant offered a chillingly vague explanation. It claimed to monitor message notification previews, bypassing direct history access while failing to clarify how exactly it harvests that data.
David Singleton of Meta Superintelligence Labs attempted to mitigate the fallout by outlining the app’s permission requirements. He noted that full functionality, including message interaction, necessitates granting the Mac application full disk access. This revelation highlights the friction between the convenience of AI-driven device management and the invasive level of system-wide permissions required to sustain it.
Comments (0)
No comments yet. Be the first!