Artificial Intelligence-Based Interactive Learning Media Integrating Social Emotional Learning and Islamic Values in Elementary Science

Abstract

This study aimed to develop an Artificial Intelligence (AI)-based interactive learning media integrated with Social Emotional Learning (SEL) and Islamic values for elementary science learning on the human circulatory system topic, and to examine its validity, practicality, and effectiveness. [Development Results] The development followed a modified Research and Development (R&D) design synthesizing the Borg and Gall and ADDIE models into seven stages: research and information collection, planning, preliminary product development, preliminary field testing, product revision, main field testing, and final product refinement. The resulting product was a web-based learning environment consisting of interactive visualization, three-dimensional organ animation, adaptive quizzes, storytelling activities, contextual reflection, and an AI-supported teacher dashboard using a rule-based adaptive system. [Validity] Expert validation involving four experts (media, instructional design, science content, and Islamic value integration) yielded an average score of 97.25%, categorized as very valid, with the highest score on instructional design (100%) and the lowest on Islamic value integration (91.50%). [Practicality] Practicality testing with 30 fifth-grade students and three teachers showed average scores of 90.90% (students) and 98% (teachers), both categorized as very practical, with the highest student response on value integration (93.10%) and the lowest on AI personalization features (89.50%). [Effectiveness] A quasi-experimental design involving 60 students (30 experimental, 30 control) demonstrated that the experimental group achieved an N-Gain of 95% (high/effective) compared with 58% (medium) in the control group. Statistical analysis confirmed significant within-group improvement (Wilcoxon: Z = -4.782; p < 0.001) and significant between-group differences (Mann-Whitney: U = 112.5; p < 0.001; effect size r = 0.65, large). SEL competence and Islamic value internalization also improved, with teacher observation rising from 41.66 to 91.04 and student self-assessment from 42.70 to 92.61. [Limitations of the Evidence] However, these findings are limited by the involvement of a single elementary school with 60 students, a relatively short intervention period, the use of rule-based rather than machine-learning-based AI, and the possible influence of teacher involvement and researcher support during implementation. Therefore, broader implementation across diverse school contexts, longer intervention periods, and more advanced AI approaches are recommended for future research.