Co-Constructing Literacy with AI: A Mixed-Methods Study of Comprehension, Motivation, and Multimodal Storytelling in Grade 4

Abstract

This exploratory, classroom-based mixed-methods study examined a five-phase, AI-supported multimodal literacy sequence implemented with 25 Grade 4 students over four weeks. The intervention combined interactive digital reading, collaborative narrative writing, AI-supported comic imagery, multimedia editing, and peer reflection. Quantitative analyses compared students’ baseline and post-intervention comprehension scores and summarized responses to a 12-item post-intervention questionnaire. Qualitative data from open-ended responses, classroom observations, and student artifacts were analyzed thematically. Students’ mean comprehension scores increased from 63.81 (SD = 9.70) at baseline to 75.37 (SD = 7.63) after the intervention, representing a significant within-group change, t(24) = 5.49, p < 0.001, dz = 1.10. Questionnaire responses were generally positive, particularly for original story writing, digital comic creation, and the addition of movement and sound. Three qualitative themes were identified: AI-supported representations as scaffolds for narrative comprehension, as catalysts for authorial identity and creative agency, and as mediators of collaborative meaning-making. Because the study used a single-group design and the comparability of the parallel assessment forms was supported through expert content review rather than large-scale psychometric equating, the findings should be interpreted as exploratory evidence of association rather than as causal effects of AI. The study nevertheless illustrates how developmentally appropriate, teacher-mediated AI activities may support multimodal literacy participation in primary education.