AI-Supported Deep Learning Model for Meaningful Character Education in Indonesian Elementary School Learning

Abstract

Elementary literacy performance in Indonesia remains below international standards, indicating the need for instructional innovation within core subjects. This study developed and evaluated an AI-supported deep learning model implemented in Bahasa Indonesia learning to enhance meaningful learning and character education in elementary schools. The research employed a Research and Development design using a sequential explanatory mixed-methods approach. The development process included needs analysis, expert validation, revision, pilot testing, and field trials involving 60 students and two teachers. Quantitative data were collected through meaningful literacy tests, character observation rubrics, learning analytics logs, and perception questionnaires, while qualitative data were obtained from interviews and classroom observations. Paired-samples t-tests showed significant improvements in conceptual understanding, critical thinking, and knowledge application. There was a notable enhancement in character indicators, particularly in the domains of empathy and responsibility. An examination of the analytics revealed a marked increase in student engagement over a period of four weeks. The integration of personalisation and analytics within the teaching of Bahasa Indonesia has been demonstrated to enhance cognitive and personal development in elementary education.