Teaching AI with games: the impact of generative AI drawing on computational thinking skills

While computational thinking (CT) is crucial for modern education, integrating artificial intelligence (AI) into learning poses challenges due to its complexity. Generative AI Drawing (GAID) offers an intuitive method for teaching AI concepts, but barriers such as restrictive access for younger students and limited instructional frameworks hinder its potential. This study proposes combining GAID with game-based learning (GBL) to create an engaging, hands-on approach for teaching CT and AI literacy. A GBL environment was developed where students designed and refined robot-control board game cards. This study involved a total of 56 sixth-grade students from two elementary schools in northern Taiwan. One group, consisting of 28 students using GAID and block-based coding to foster AI literacy, was assigned as the experimental group. The other group, consisting of 28 students who did not employ GAID but relied only on a Google search engine in their learning, was assigned as the control group. The results showed that the experimental group enhanced algorithmic thinking and AI literacy, particularly in the “Create AI” dimension, compared to the control group. However, the control group excelled in CT concept mastery, suggesting that the beginners need conventional learning first to better support foundational CT learning. Consequently, a balanced educational approach blending automated tools like GAID with exploratory, project-based activities is recommended to maximize learning outcomes.