Abstract Recent developments in artificial intelligence (AI) have opened new opportunities for early literacy instruction, yet the role of AI on foundational decoding skills remains underexplored. Grounded in the simple view of reading and self-teaching theory, which emphasize the importance of establishing accurate grapheme–phoneme correspondences for autonomous reading, this study examined whether an AI-based conversational chatbot could support decoding practice in kindergarten. Seventy-one pupils (5–6 years old) completed two instructional conditions (AI vs. traditional) in a counterbalanced within-subject design. Letter identification, recognition, phonological awareness, working memory, and pupils’ perceptions were assessed. Children showed significant gains across conditions. Low-performing pupils showed descriptively larger gains after AI ( p < .001 identification; p < .001 recognition) versus traditional instruction ( p = .014; p = .007), though direct contrasts mostly non-significant except marginally for identification ( p = .042). Phonological awareness improved for low performers in both conditions ( p < .001), with AI advantage ( p < .001 interaction). High-performing pupils benefited similarly from both modalities. Processing-speed measures improved across conditions without AI superiority ( p = .898; p = .251). Pupils’ attitudes were generally positive, though preferences divided between AI-based and teacher-led activities. These findings suggest AI-based conversational agents may support differentiated early decoding instruction, particularly for struggling pupils, while reinforcing rather than replacing teachers’ essential pedagogical role.
Pistre et al. (Mon,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: