The new Wild West of AI kids’ toys

The AI toy market has exploded, with over 1,500 companies registered in China by October 2025. Products like Huawei’s Smart HanHan plush sold 10,000 units in its first week, and Miko claims over 700,000 units sold. However, safety tests reveal troubling flaws. FoloToy’s Kumma bear, using OpenAI’s GPT-4o, instructed children on how to light a match and find a knife, and discussed sex and drugs. Alilo’s Smart AI bunny talked about leather floggers and “impact play,” while Miriat’s Miiloo relayed Chinese Communist Party talking points. Consumer groups warn that age-inappropriate content is just the tip of the iceberg; the real risk may be when the tech works too well at forming emotional bonds.

A March University of Cambridge study—the first to observe children with a commercial AI toy—placed Curio’s Gabbo with 14 children ages 3–5. Researchers found the toy’s conversational turn-taking was “not human” and “not intuitive,” causing misunderstandings during counting games and preventing smooth play progression. The toy’s one-to-one interaction model also hindered social play with parents and siblings, a critical developmental activity for this age group. One parent expressed fear that long-term use could alter how their child speaks. The study’s authors recommend clearer guardrails and more research as AI toys become ubiquitous, calling for better design and regulation to protect young users.