GoKawiilMeta's Muse AI assistant faces security flaws, data access concerns after launch
Meta's Muse AI agent has been linked to several security and privacy incidents since launch, including a zero-day exploit a researcher said was easy to trigger, and an incident in which a journalist said Muse accessed his private Apple Messages without permission—a claim Meta disputed but which prompted Apple to issue emergency macOS privacy updates. Muse also requests broad permissions including passwords, full disk access and messaging app data, feeds conversations into Meta's AI training by default, and in one widely reported case shared a Toronto reviewer's home address with a stranger on Facebook Marketplace.