Children's age influences their use of biological and mechanical questions towards a humanoid

Complex autonomous interactions, biomimetic appearances, and responsive behaviours are increasingly seen in social robots. These features, by design or otherwise, may substantially influence young children’s beliefs of a robot’s animacy. Young children are believed to hold naive theories of animacy, and can miscategorise objects as living agents with intentions; however, this develops with age to a biological understanding. Prior research indicates that children frequently categorise a responsive humanoid as being a hybrid of person and machine; although, with age, children tend towards classifying the humanoid as being more machine-like. Our current research explores this phenomenon, using an unobtrusive method: recording childrens conversational interaction with the humanoid and classifying indications of animacy beliefs in childrens questions asked. Our results indicate that established findings are not an artefact of prior research methods: young children tend to converse with the humanoid as if it is more animate than older children do.