Friendship with a robot: Children's perception of similarity between a robot's physical and virtual embodiment that supports diabetes self-management.

OBJECTIVE: The PAL project develops a conversational agent with a physical (robot) and virtual (avatar) embodiment to support diabetes self-management of children ubiquitously. This paper assesses 1) the effect of perceived similarity between robot and avatar on children's' friendship towards the avatar, and 2) the effect of this friendship on usability of a self-management application containing the avatar (a) and children's motivation to play with it (b). METHODS: During a four-day diabetes camp in the Netherlands, 21 children participated in interactions with both agent embodiments. Questionnaires measured perceived similarity, friendship, motivation to play with the app and its usability. RESULTS: Children felt stronger friendship towards the physical robot than towards the avatar. The more children perceived the robot and its avatar as the same agency, the stronger their friendship with the avatar was. The stronger their friendship with the avatar, the more they were motivated to play with the app and the higher the app scored on usability. CONCLUSION: The combination of physical and virtual embodiments seems to provide a unique opportunity for building ubiquitous long-term child-agent friendships. PRACTICE IMPLICATIONS: an avatar complementing a physical robot in health care could increase children's motivation and adherence to use self-management support systems.

Using a robot to personalise health education for children with diabetes type 1: a pilot study.

OBJECTIVE: Assess the effects of personalised robot behaviours on the enjoyment and motivation of children (8-12) with diabetes, and on their acquisition of health knowledge, in educational play. METHODS: Children (N=5) played diabetes quizzes against a personal or neutral robot on three occasions: once at the clinic, twice at home. The personal robot asked them about their names, sports and favourite colours, referred to these data during the interaction, and engaged in small talk. Fun, motivation and diabetes knowledge was measured. Child-robot interaction was observed. RESULTS: Children said the robot and quiz were fun, but this appreciation declined over time. With the personal robot, the children looked more at the robot and spoke more. The children mimicked the robot. Finally, an increase in knowledge about diabetes was observed. CONCLUSION: The study provides strong indication for how a personal robot can help children to improve health literacy in an enjoyable way. Children mimic the robot. When the robot is personal, they follow suit. Our results are positive and establish a good foundation for further development and testing in a larger study. PRACTICE IMPLICATIONS: Using a robot in health care could contribute to self-management in children and help them to cope with their illness.