Week 2

A is for Artificial Intelligence

The Impact of Artificial Intelligence Activities on Young Children’s Perceptions of Robots

Randi Williams                                                                 HaeWon Park                                                      Cynthia Breazeal

The authors have developed an AI platform PopBots for preschool children. This platform consists of a social robot toolkit which is a programming interface on a tablet computer. Three hands-on activities with assessments for young children were built to explore machine learning and reasoning algorithms. Children use the tablet with block-based interface which is picture-based, built on Scratch blocks to program and train the robot for each AI activity.

Three AI concepts: Knowledge-based Systems, Supervised Machine Learning and Generative AI were each packaged in a child-friendly activity. Assessment was conducted for each activity as a set of multiple-choice questions to probe what children understood about each AI concept. Assessments were verified in the pilot tests and children’s performance was compared with their demonstrated understanding interviews. Through these assessments like Perception of Robots Questionnaire and Theory of Mind Assessment, children’s learning outcomes were analyzed. Assessments were done with everyone in the classroom and children did not share or discuss their answers until after everyone had answered. The quantitative data were collected through children’s responses from the Questionnaires in which children’s questions and observations and their understanding of their activities along with the reasoning about their responses was recorded for further analysis.

A performance on Theory of Mind Assessment was analyzed and it is observed that the knowledge access question was the easiest for children to get correct at 75 %. Children’s score was graded when a question is only correct if a child gets both the target and control question correct. The false belief questions proved to be more difficult with both correct at 55 % of correct answers.