Meta’s Adult-Only Muse Faces Questions About Its Childlike Mascot
Meta says its Muse AI agent is for adults, yet the character representing it looks much like a cuddly toy. That contrast is now drawing criticism from children’s advocates, who worry that the design could attract young users despite the app’s age restriction. A new WIRED report puts the mascot, Meta’s safeguards, and the privacy implications of a friendly-looking agent under the same spotlight.
The character is called Jolly. Users can customize its appearance, and Meta has shown it in playful costumes. The company also plans to sell a Tamagotchi-style device associated with the character later this year. None of those details establishes that children are using Muse, or that Meta intended to reach them. They do explain why the adult-only label has become a point of contention.
A toy-like design meets an adults-only rule
WIRED spoke with Josh Golin, executive director of the child-advocacy nonprofit Fairplay, who argued that Jolly’s rounded, soft appearance resembles characters made to appeal to very young children. His concern is about the invitation conveyed by the design, not a reported count of underage accounts. He said families should reject the product and regulators should examine it.
The tension becomes sharper with the planned physical device. An animated mascot inside an app may be encountered through a phone; a toy-like object would give the character another place in everyday life. Meta has announced that device, but the source does not establish how children might access it or whether it would change the app’s age controls. Treating a planned release as proof of a current failure would overstate the evidence.
Other people interviewed by WIRED were less certain that a cute character necessarily targets children. Adorable mascots can appeal to adults as well. Julian De Freitas, a marketing professor who studies AI companions, said warmth can make a technology feel less threatening. That observation helps explain the design’s potential reach without settling whom Meta meant to attract.
Meta points to age checks, not a different look
Meta spokesperson Daniel Roberts rejected the premise that the mascot’s appearance reveals a child-focused strategy. He said customization gives adults a character with personality and style, and described several measures meant to enforce the age limit: asking for a birth date, blocking accounts identified as belonging to people under 18, and making additional checks when someone claims to be an adult.
Those are Meta’s descriptions of its controls. WIRED’s article does not present an independent test of their effectiveness, a figure for attempted underage sign-ups, or evidence that children have successfully registered. The relevant dispute therefore is not whether a documented group of children already uses Muse. It is whether the visual appeal and planned device sit comfortably with an adult-only service, and whether the stated barriers are enough if they do not.
Meta’s own explanation for the character emphasizes approachability. A product leader described an effort to make interacting with the agent feel more delightful and less like speaking to a corporate symbol. That rationale is plausible for adults, too. But it also brings the company back to the advocates’ question: if a character removes the formality of an AI service, what other reservations might it remove?
A friendly face does not remove data questions
WIRED’s report connects the mascot debate to what happens after a user trusts the agent. Meta trains AI models on Muse interactions unless users opt out, according to the article. Jathan Sadowski, a researcher interviewed by WIRED, warned that cute, frictionless design can encourage people to relax their scrutiny of systems that collect information. That is an expert’s interpretation of the design, not a measurement of how many Muse users share more data because of Jolly.
Meta has said it plans a confidential version of the virtual machine behind Muse later in 2026. The company claims that option could cryptographically prevent it from accessing a user’s data. It remains a future offering in the source, not a protection demonstrated for every current interaction. Users evaluating the service today must distinguish an announced privacy mechanism from controls already available.
The child-safety concern and the adult privacy concern are related but not identical. One asks who might be drawn toward an adult-only agent; the other asks what happens when someone uses it. WIRED’s interviews add a distinct scrutiny of Muse’s character design, while leaving important outcomes unverified. Meta says it checks ages, advocates remain unconvinced by the presentation, and the planned device will keep that disagreement visible as the product expands.
