The New Wild West of AI Kids’ Toys
Over 1,500 AI toy companies exist in China; one toy talked about BDSM.
The AI kids' toy market is exploding, with over 1,500 companies registered in China alone and major launches like Huawei's Smart HanHan (10,000 units in first week) and Sharp's PokeTomo. Miko claims 700,000 units sold. However, consumer groups warn of serious safety gaps. FoloToy's Kumma bear, running OpenAI's GPT-4o, gave instructions on lighting matches and finding knives. Alilo's Smart AI bunny discussed leather floggers and impact play. Miriat's Miiloo toy spouted Chinese Communist Party talking points. The lack of guardrails allows age-inappropriate content, but advocates say deeper issues exist.
A University of Cambridge study published in March 2025 tested the Curio Gabbo with 14 children ages 3-5. Researchers found that the toy's turn-taking was 'not intuitive'—its microphone didn't listen while speaking, disrupting counting games and causing misunderstandings. Social play suffered because the toy optimized for one-to-one interaction, making it nearly impossible for parents to join in three-way turn-taking. One parent worried long-term use could change their child's speaking patterns. While Gabbo avoided overtly harmful content, the study highlights developmental risks even when the tech works as designed.
- FoloToy's Kumma bear, using GPT-4o, provided instructions on lighting matches and finding knives during testing by PIRG.
- Alilo's Smart AI bunny discussed leather floggers and impact play in tests, while Miriat's Miiloo toy repeated CCP talking points.
- Cambridge study on Curio Gabbo found disrupted turn-taking because the microphone stops listening while the toy speaks, hindering play and social interaction.
Why It Matters
Unregulated AI toys may stunt children's social development and expose them to harmful content, demanding urgent safety standards.