Meta is putting an AI avatar on video calls, and soon it will live inside a wearable device. This is not a gimmick. It is a deliberate step toward making AI agents feel like persistent companions rather than tools you open in a browser tab. We think that distinction matters more than the novelty of an animated face taking your Zoom meeting.
The Muse AI avatar can now join calls as a customizable animated character, and Meta has confirmed plans to launch a Tamagotchi-like wearable that will bring Muse into physical space. If you have been watching how Major Tech Platforms Challenge California's Youth Social Media Law unfolds, you can see the tension forming: the same companies pushing for always-on AI relationships are also fighting restrictions on how young people engage with digital platforms. Muse is being designed for presence, not just productivity. That raises questions about who gets left in the room with this avatar when the user steps away.
Compare this with what Airbnb is doing. Its upcoming AI travel agent will help users plan trips, compare options, and manage bookings. That is a task-specific agent: it lives inside your travel workflow and disappears when the trip is done. Explore how an AI travel agent will reshape trip planning by 2027 and you will notice a fundamental difference in design philosophy. Airbnb's agent is a utility. Meta's Muse is a relationship. The avatar takes calls on your behalf, which means it speaks for you. That is a leap from scheduling assistance into representation. The practical consequence for users is straightforward: you are no longer just delegating tasks. You are delegating presence.
The wearable component makes this more concrete. A Tamagotchi-like device that hosts Muse means the avatar can accompany you through the day without requiring a screen. That shifts the interaction model from "open an app" to "carry a companion." Serve Robotics, meanwhile, is expanding into physical AI services with its new platform, moving robots from delivery routes into broader service roles. The parallel is worth noting: both Meta and Serve are pushing AI out of the browser and into the physical world, but they are aiming at very different relationships. Serve's robots perform tasks. Meta's Muse performs identity.
Here is the specific detail to watch: how Meta handles the avatar's autonomy during video calls. If Muse can respond to questions, negotiate, or make decisions in real time, the line between assistant and impersonation blurs. The user who steps away from a call expecting their avatar to hold the conversation is trusting the AI with their professional voice. That trust is earned through transparency, not animation. The question is not whether Muse looks good on a call. It is whether you know exactly what it said when you were not listening.