AI-Assisted Content Adaptation for Personalized Learning in a Multimodal Narrated Digital Picture Book Prototype for Early Childhood Marine Environmental Literacy

Abstract

Marine environmental literacy in early childhood requires learning media that can convey ecological issues in a simple, visual way and align with the characteristics of child development. Topics such as marine pollution, changing ecosystem conditions, and human responsibility for the environment are often abstract for 5–7-year-olds, so an approach is needed that not only relies on text but also incorporates visuals, narratives, sounds, simple motion, and children's responses. This research aims to develop a set of multimodal design criteria for a narrated digital picture book to support early childhood marine environmental literacy. The design process employs AI to generate personalized narratives based on children's emotional responses and to develop guided questions and answers that correspond to the difficulty level of the intended learning objectives. This study adopts a qualitative, practice-based design research approach, with AI supporting the content creation process. Preliminary data were collected through a literature review, visual observations, and assessments of children's responses to ocean-themed picture books during read-aloud sessions. The findings indicate that children were more engaged when the learning media featured clear and easily identifiable marine animal characters, recognizable visual elements, simple storylines, and guided questions. These findings informed the development of a digital prototype incorporating marine characters, narrated audio, environmental soundscapes, simple animations, changes in visual atmosphere, and adaptable AI learning support. A limitation of this study is the absence of empirical validation of the final prototype with end users.