This study aims to systematically analyze scientific literature regarding the application of the ADDIE model in developing interactive multimedia for Coding and AI subjects at the primary school level, identifying research trends, best practices, and gaps requiring further academic attention. The study adopts a Semantic Literature Review (SLR) approach incorporating systematic mapping and semantic analysis. The PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) protocol was applied to select articles from Scopus, Web of Science, ERIC, and IEEE Xplore databases. A total of 47 Scopus-indexed journal articles (2015–2025) were analyzed using thematic synthesis and conceptual mapping techniques. The analysis reveals a significant increase in the application of the ADDIE model for interactive multimedia development, with a primary focus on the Development phase, which integrates AI tools for content acceleration. Key findings include: (1) a dominance of asynchronous e-learning approaches (70%) in implementation; (2) the integration of multimedia presentations through short films (35%) as a best practice; (3) a critical need for a comprehensive AI literacy framework for educators; and (4) a research gap in long-term evaluations of learning effectiveness. The ADDIE model proves effective as a systematic framework for developing Coding and AI learning multimedia in primary schools; however, it requires adaptation through generative AI integration and agile approaches to remain responsive to rapid technological advancements. Future research should focus on developing valid and reliable assessment tools to measure AI literacy in primary school students.