Cute AI agents raise a serious question: who are they designed to appeal to?

Meta’s Muse assistant and OpenAI’s Dots use friendly, character-like designs, prompting children’s safety advocates to question whether products described as adult-focused could also attract younger users. The debate is not just about a mascot’s face: it touches product design, age safeguards and where child-protection rules apply to general-purpose AI.

Watch Desk analysis

What happened

Bloomberg Law reports that Meta’s Muse defaults to a furry character called Jolly, while OpenAI’s Dots are soft, colourful shapes. The companies say the products are intended for adults. Meta says Muse users must provide their date of birth, that it blocks people it detects may be under 18 from creating accounts, and that it carries out additional age checks. OpenAI says Dots are not available to under-18s and use a neutral grey circle by default.

Children’s advocates quoted by Bloomberg Law argue that the designs could make the assistants feel like toys or trusted companions. A technology-policy expert interviewed by the publication offers a counterpoint: a character-like appearance alone may not mean a general-use product is legally directed at children. The article also describes differing US rules: existing federal COPPA protections concern the collection of children’s personal information, while some newer state measures focus on physical toys with built-in chatbots. How those laws apply to these assistants is not settled in the account.

Why it matters

An AI assistant’s appearance can shape who feels invited to use it, but design is only one part of the child-safety question. Age checks, what an assistant can say or do, and how it handles children’s data matter too. The gap between an adult-use intention and a child’s perception is exactly where product teams and regulators may find themselves arguing over the fine print.

Our read

A friendly interface is not proof that a product targets children, and a label saying “for adults” is not a complete safety plan. The useful test is whether the safeguards work for the people who may encounter the product, not only the audience its makers had in mind. Cute is a design choice; responsibility comes with it.

What to watch

• Whether Meta or OpenAI changes character defaults, access controls or age checks.
• How state and federal rules are applied to general-purpose AI assistants, not just physical toys.
• Whether future evidence shows who is using these agents and how effectively age safeguards work.

Discussion spark: Should AI companies have to design general-purpose assistants to be less appealing to children, or should effective age checks and parental choice be enough?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.