Young children's understanding of artificial intelligence: A draw‐a‐picture analysis

With the increasingly ubiquitous presence of artificial intelligence (AI) technology in digital society, educators appeal to foster the public's general knowledge about AI from young children. Studies have been conducted to create learning activities for children's development of AI literacy, but seldom did they investigate the resources children have prior to instruction. The purpose of this study is to explore young children's preconceptions of AI and the sources of their viewpoints. A total of 261 students aged 7 to 9 from a public elementary school in a Southern Chinese city were recruited to finish the draw‐a‐picture task, and 32 students participated in the follow‐up sharing in the classrooms. Through the drawing analysis, we identified five categories to describe young children's impressions of AI, including the appearance of AI, tasks handled by AI, environment, human–AI relationship, and feelings and attitudes towards AI. Moreover, our analysis revealed that everyday technology and media act as two main sources of young children's preconceptions of AI. This research offers methodological and empirical contributions to AI literacy research and practical insights for AI production developers, policymakers and curriculum designers. Practitioner notes What is already known about this topic AI literacy, as an important part of digital literacy in the twenty‐first century, has drawn attention from researchers. Some studies designed AI literacy courses and explored their influence on young children's understanding of AI. Children have informal ideas and beliefs about AI prior to instruction, which is tricky to catch. What this paper adds By draw‐a‐picture analysis, 5 categories and 28 subcategories were identified to describe young children's preconceptions of AI technology. For AI appearances, children put animal‐like features in nearly the same important position as human‐like characteristics. For AI appearances, performed tasks and environment, children consider AI as not only mimicking humans but also an extension and surpassing of human intelligence. Human–AI relationships perceived by children are identified as teammate/partnership, commanders–subordinates, master–server and substitute relationships. Children's positive attitudes towards AI are prominent, but they have shown the potential to understand possible negative influences of AI. Two sources of young children's AI preconceptions, everyday technology and media, were revealed. Implications for practice and/or policy Drawing is an age‐appropriate assessment approach for investigating children's understanding of AI conceptions. For AI developers, children's tendency to overestimate AI capabilities presents a challenge. AI products may be designed with human‐like, animal‐like or tool‐like appearances to reduce children's cognitive load; children's socio‐cultural resources can be integrated into user interface design to enhance their engagement; and drawing‐based participatory design workshops can be conducted to incorporate children's voices into the design of AI applications. Policymakers should develop regulations to ensure the responsible design of AI for children; provide guidance to foster children's media literacy; and incorporate AI literacy education into the formal curriculum. Curriculum designers can utilize situated learning strategies and culturally responsive pedagogy to engage children in AI literacy education. Curriculum designers should incorporate instructions about the limitations of AI systems, ethical issues AI might introduce, and the differences between fiction–fact and AI–other technology into AI literacy education.