The development of Artificial Intelligence (AI) has encouraged the emergence of various learning media innovations with the potential to improve the quality of education in elementary schools. However, studies specifically discussing AI-based learning media at the elementary education level remain limited and dispersed across different publications. This study aims to identify the characteristics of research, types of AI-based learning media, their benefits, implementation challenges, and development trends in elementary education. The study employed a Systematic Literature Review (SLR) method based on the PRISMA 2020 guidelines. The population consisted of scientific articles discussing AI-based learning media in elementary education. The sample comprised ten articles that met the inclusion criteria and were retrieved through Google Scholar. Data were analyzed using descriptive and thematic analysis techniques to identify patterns and key findings across the selected studies. The results indicate that the most commonly used AI-based learning media include Adaptive Learning Systems, Intelligent Tutoring Systems, Educational Chat bots, Generative AI, Learning Analytics, and Automated Assessment. These technologies support more personalized, adaptive, interactive, and student-centered learning experiences. Nevertheless, their implementation still faces several challenges, including limited infrastructure, teachers’ readiness, the digital divide, and issues related to ethics and data security. In conclusion, AI-based learning media have significant potential to enhance the quality of learning in elementary schools when supported by adequate resources and appropriate implementation strategies.