Meta's Muse Agent Wants Your Data — And Everyone Else's

Meta's Muse AI assistant hit No. 1 on Apple's App Store and accumulated millions of downloads within a month of release, according to The Verge — and it is asking users to hand over data that belongs to other people.

Muse belongs to a category of products The Verge describes as "AI personal assistants." Unlike a conventional chatbot that answers questions you type, an agent (software that acts on your behalf across apps and services) needs continuous access to your messages, contacts, calendar, and context to function. Muse sits squarely in that category. The Verge's reporting frames the product as the latest step in a broader consumer tech trend: "gadgets designed to spy on everything and everyone around you, regardless of whether or not anyone else actually consented to it."

That sentence is the crux of the story. The consent problem here does not concern the user, who clicked through the terms of service. It concerns everyone the user knows.

What does Muse actually ask for?

According to The Verge, Muse asks users to contribute not only their own data but also information about "friends, family, the person you went on a Hinge date with five years ago — and virtually anyone else you've ever come in contact with."

The mechanism is straightforward. An AI agent's usefulness scales with the data it can see. To plan a dinner, it needs to know who your friends are. To manage your calendar, it needs to know who you are meeting. To draft messages, it needs a model of your relationships. Each of those functions pulls in third parties — people who never agreed to anything.

The Verge's reporting also notes Muse's presentation, describing what it calls its "creepy virtual avatars" in its assessment of the product's design choices. The tone of the coverage is pointed, and the framing matters: this is not a story about one app's settings menu. It is a story about a structural shift in what consumer AI products expect from their surroundings.

Why millions of downloads change the stakes

A consent debate about a niche product would stay niche. Muse is not niche. The app reached the top of the App Store charts within roughly a month of launch, with millions of downloads.

Those numbers convert an ethical question into a practical one. If a single user opts into an always-on agent, the exposure is limited to that user's circle. If tens of millions of users do it in the first month, the population of people whose information flows into Meta's systems without their agreement grows accordingly — friends, family, former dates, colleagues, and anyone else in a contact list.

This is the bystander problem in its purest form. The Verge's headline — "Friends don't let friends feed their data to AI agents" — is not a joke about etiquette. It is a literal description of the situation: the person deciding what data enters the system is not the person the data describes.

Where does this fit in the market?

The Verge situates Muse within a wider pattern it has tracked previously, linking the assistant to an existing wave of consumer devices and wearables that capture people in the user's environment — a category the outlet has covered under the label of surveillance-creep tech, including smart glasses and connected watches.

The common thread across those products and Muse is asymmetry. One party buys the device or downloads the app. Everyone else in the room, the chat, or the contact list gets captured as a side effect.

For Meta, the commercial logic is clear. Agents trained on richer personal context deliver more useful output, and more useful output drives adoption — which the App Store ranking suggests is already happening. What the company has not yet answered, at least in the material available from this reporting, is what recourse exists for the people in the data who never signed up.

What is not in the story

The Verge's piece, as excerpted here, does not specify Muse's exact data-sharing terms, retention policies, or opt-out mechanisms for third parties. That absence is itself worth flagging. A product that ingests information about millions of non-users, at the scale implied by these download figures, raises questions the launch cycle has so far outrun.

Until those mechanics are documented, the only verified facts are these: Muse launched last month, it is downloaded millions of times, it holds the No. 1 spot on the App Store, and its own marketing premise is that your social graph — not just your data — is the input. The Verge's full report, which covers the product in greater detail, can be read at theverge.com.

The takeaway in plain words: the fastest-growing AI agent on the App Store treats your consent as covering everyone you know, and the industry has not decided whether that is acceptable.