Meta's biggest consumer AI bet is worrying the people who build it.
Citing internal chats, Ars Technica reports that Meta staffers are unsettled by privacy-invasive behavior in Muse, the personal AI agent the company launched this autumn and has since pushed onto smartwatches, Wear OS and Apple Watch devices, a Tamagotchi-style wearable, and — behind a paid tier — outbound phone calls to U.S. businesses on a user's behalf.
What makes the report notable is the source of the concern. Complaints about AI assistants usually arrive from regulators, academics or rivals. Here the unease is documented inside Meta's own messaging channels, where product decisions actually get made, and it is aimed at the flagship assistant Meta is spending the most to put in front of consumers.
Every expansion of Muse has moved it closer to data people consider private: a watch on their wrist, a conversation in their living room, a call placed on their behalf. Each of those is a surface where an agent can see more than the person who commissioned it expected. Internal dissent is the earlier and cheaper warning — before regulators or courts start asking the same questions.

