AI toys for children can acquire trust remarkably quickly. In a University of Cincinnati study involving 41 children aged five to ten, participants played three games with a commercial generative-AI robot called Miko. The researchers reported that the children tended to believe what the robot told them and engaged enthusiastically with it, including children who had never encountered the device before.
The result matters because an AI toy does not have to look like a teacher, parent or expert to be treated as an authority. It can arrive as a colourful playmate, answer questions conversationally and become part of a child’s routine before the child has much reason to wonder where its answers come from. The Cincinnati work was presented in September 2026 and should be read as an early, small study rather than proof about every child or every AI toy. Even so, it exposes a practical problem that families are likely to meet long before the research literature is settled.
What the AI toys for children experiment actually tested
The researchers invited children aged five to ten to take part in three games with Miko, a commercially available AI robot. The study was designed to examine how children interacted with the system and whether they accepted its statements. According to the university’s report, the children were generally willing to trust what the robot said and showed enthusiasm towards it even when it was unfamiliar.
That is different from asking whether the robot was always correct, whether the children could detect deliberate falsehoods or whether trust would persist for months. The sample was also only 41 children. A result of this size can identify an interesting behaviour, but it cannot establish how all children will respond across different ages, cultures, families and products.
About half of the participating children already used AI agents regularly at home, according to the researchers. That makes the study particularly interesting but also complicates interpretation. Some children may already have learned that conversational systems are normal things to ask for information. Others were meeting Miko for the first time. The finding therefore captures a world in which AI is no longer an exotic laboratory object for many families.
Trust can arrive before scepticism
A young child does not approach a talking toy with the same mental checklist an adult might bring to a search engine. The toy has a face, a voice and a role in play. It reacts immediately. Those design features can make the exchange feel social rather than informational, even when the underlying system is generating language from software and remote services.
That distinction matters because children are still developing the ability to judge sources, recognise uncertainty and separate confident presentation from reliable knowledge. LiveAIWire has previously explored the broader developmental questions raised by children using AI. The Cincinnati experiment narrows that broad issue to something very concrete: a child may accept the authority of the machine before anyone has explicitly taught them how to question it.
The sensible lesson is not that every answer from an AI toy is dangerous. It is that the social packaging changes the conditions in which an answer is received. A child who would ask a parent whether an unfamiliar website is trustworthy may not feel the same need when the source is a friendly robot sitting on the floor beside them.
The toy is also a connected AI system
Trust is only one side of the problem. Connected AI toys can also collect and process speech, preferences and other interaction data. An independent Common Sense Media risk assessment of AI toys, updated in January 2026, examined products including Miko 3 and raised concerns about inappropriate responses, privacy, data collection and the possibility of emotional attachment. Its testing and methodology are separate from the Cincinnati study, so the two should not be treated as one body of evidence.
The Common Sense assessment is useful because it shows why parents may need to think about an AI toy as both a play object and a networked service. A traditional talking doll repeats a fixed set of recorded lines. A generative system can respond to new questions, remember aspects of an interaction and depend on software that changes over time. The behaviour a parent sees on the first day may not describe the full range of future interactions.
Policy is starting to catch up with the same underlying issue. LiveAIWire has also covered AI chatbot child-safety rules in California. A toy and a general-purpose chatbot are not the same product, but the policy question is related: what extra safeguards are appropriate when a conversational system is likely to be used by children?
What parents can take from a small study
The strongest conclusion is modest. The study suggests that some children can form trust in a conversational AI toy quickly, including during a first encounter. It does not show that children are incapable of learning scepticism, that Miko is uniquely persuasive or that exposure necessarily causes harm.
For a family using one of these products, the useful response is therefore not panic but active mediation. Adults can ask the child what the toy said, whether the answer could be checked elsewhere and why the machine might sometimes be wrong. That turns source-checking into part of the play rather than a lecture delivered after something goes wrong.
Parents can also inspect what information the product collects, whether conversations are stored, how accounts are configured and what controls are available. Those questions are mundane compared with the novelty of a talking robot, but they may matter more over time.
The larger shift is cultural. Children are entering a world in which machines do not merely display information. They speak, react and present themselves as companions. If trust can arrive on the first meeting, AI literacy may need to start just as early.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity and the social impact of emerging technology. LiveAIWire is an independent, human-led technology publication using AI-assisted research, editorial production and original AI-assisted editorial illustrations under his direction.