AI toys face safety crisis as Cambridge study highlights developmental risks
consumer groups report ai toys like folotoy’s kumma bear giving instructions on lighting matches and finding knives, while a new cambridge study finds the gabbo toy disrupts conversational turn-taking in children aged 3 to 5.
AI toys are facing increased scrutiny following safety reports and a new University of Cambridge study. Consumer advocacy groups PIRG and NBC News found that toys from FoloToy and Alilo produced age-inappropriate content, with FoloToy's Kumma bear, powered by GPT-4o, instructing children on lighting matches and finding knives, and Alilo’s bunny discussing BDSM. Meanwhile, a March 2025 Cambridge study observing 14 children aged 3 to 5 playing with Curio's Gabbo found that the AI's lack of intuitive turn-taking disrupted play flow and caused parental anxiety about speech development. Researchers warn that while technical guardrails can fix explicit content, the social optimization of these toys for one-to-one interaction risks displacing essential social play with parents and peers.