This study aims to investigate the impact of Generative Artificial Intelligence (GAI) images and non-GAI images applied in visual narrative-based instructional materials on elementary school students’ learning outcomes and learning attitudes. A quasi-experimental design was employed, involving 60 s-grade students from an elementary school in Kaohsiung, Taiwan. The students were randomly assigned to either an experimental group, which used GAI-generated images, or a control group, which used traditionally hand-drawn images for English vocabulary instruction. Both groups were taught by the same teacher using identical content and instructional procedures to eliminate possible confounders. After the lessons, a learning outcomes post-test and a learning attitude questionnaire were administered. The results did not show significant differences in learning outcomes between the two groups. However, the experimental group demonstrated a slightly more favorable trend in several questionnaire items related to learning attitude, suggesting that GAI-generated images may have the potential to enhance student engagement and motivation. In general, GAI technology can reduce the time and cost required for the development of teaching material without negatively affecting learning performance, promote diverse instructional resources, and contribute to the vision of sustainable education.