Delving into primary students’ conceptions of artificial intelligence learning: A drawing-based epistemic network analysis

Artificial Intelligence (AI) education in primary schools has received a great deal of attention globally, and it is thus important to investigate primary school students’ perceptions and understanding of AI learning. Therefore, in this study, 673 drawings of conceptions of AI learning by third to sixth grade students were collected. Firstly, a drawing analysis approach was used to code the drawings into five categories of 25 elements. Then, the epistemic network analysis approach was used to visualize the structure of the relationships between the elements within the conceptions of AI learning. The study found that (1) primary school students generally held positive attitudes toward AI learning and understood AI learning as programming or robot programming learning; (2) higher grade students showed more constructive learning concepts; and (3) girls had significantly more positive emotions and attitudes toward AI learning than boys, but had less participation in manipulating robotic human activities than boys. The over-reliance on programming and robotics in elementary AI education, as well as gender differences in elementary AI learning need more attention from educators.