Artificial Intelligence in Learning Multimedia for Elementary Education: A Systematic Literature Review

Abstract

This study aims to map research trends, identify types of artificial intelligence (AI) technologies employed, analyse their effectiveness on student learning outcomes, and identify research gaps in the utilisation of AI-based learning multimedia at the elementary school level. A systematic literature review (SLR) guided by the PRISMA 2020 protocol was conducted across three internationally reputable databases Scopus, Web of Science, and ERIC yielding 29 selected articles from 15 countries (2021–2025) through a rigorous selection process (κ = 0.83). Findings reveal an accelerating growth in publications dominated by East and Southeast Asian countries. Seven categories of AI technology were mapped: adaptive learning systems, augmented reality, conversational AI agents, AI-powered digital games, virtual reality, learning analytics, and generative AI. Cross-study evidence consistently demonstrates the effectiveness of AI-integrated multimedia in improving academic achievement, motivation, and learning personalisation. Five critical research gaps were identified: the scarcity of longitudinal studies, limited research in non-WEIRD contexts, insufficient attention to ethics and data privacy, minimal AI integration in non-core subjects, and the absence of comprehensive teacher professional development frameworks for AI implementation.