Challenges in Implementing the Deep Learning Approach in Indonesian Elementary Schools
Published 2026-08-31
Keywords
- learning model,
- deep learning,
- elementary school
Abstract
Deep learning pedagogy, characterized by mindful, meaningful, and joyful learning rather than artificial intelligence, is being promoted in Indonesia to move elementary education beyond rote learning. However, evidence concerning its implementation challenges remains fragmented. This structured literature review aimed to identify and synthesize the challenges, influencing factors, response strategies, and implications of implementing deep learning pedagogy in Indonesian elementary schools. Google Scholar, GARUDA, and Dimensions were searched for empirical journal articles published in Indonesian or English between January 2016 and December 2025. The review followed PRISMA 2020, while methodological quality was appraised using the Mixed Methods Appraisal Tool. Seven articles met the eligibility criteria. All were published in 2025; four employed quantitative survey designs and three used qualitative approaches. Thematic synthesis identified five interrelated challenge categories: conceptual, pedagogical, technological, institutional, and classroom-related. Institutional and technological challenges were each reported in five articles, while conceptual and pedagogical challenges each appeared in four. Key barriers included uneven conceptual understanding, confusion with project-based learning, moderate teacher readiness, limited training and mentoring, inadequate or underused technology, insufficient school support, high teacher–student ratios, and student diversity. Diagnostic assessment, cooperative learning, reflective feedback, and contextual media emerged as practical responses. Effective implementation requires sustained professional development, supportive leadership, professional learning communities, aligned policies, and context-sensitive resources. The evidence remains limited by small heterogeneous samples, concentrated geographical coverage, and the predominance of teacher perspectives. Future multisite, longitudinal, mixed-methods, and intervention studies should examine implementation fidelity and objective student outcomes.