PERAN AI DALAM DUNIA PENDIDIKAN DI SEKOLAH DASAR

Abstract

This study aimed to compare the contributions and limitations of artificial intelligence in elementary schools. The method used was a systematic literature review with light bibliometric analysis of articles published between 2020 and 2025. The review focused on differentiated learning, assessment and formative feedback, explainable AI, human-in-the-loop mechanisms, and data governance. Following a staged screening process based on PRISMA guidelines, 72 articles were analyzed thematically and mapped using keyword co-occurrence. The results indicate that most studies reported improvements in literacy and numeracy when artificial intelligence was used to support instructional strategies, with a moderate average effect size. Automated feedback systems reduced teachers’ routine workload by approximately 24 minutes per week and improved the frequency and accuracy of responses. While the detection of misconceptions was considered adequate, the level of agreement between AI and teacher assessment ranged from moderate to high.However, several challenges remain, including uneven fairness audits, limited teacher control features, insufficient system transparency, and low levels of open science practices. In conclusion, artificial intelligence is most effective when it reinforces teachers’ pedagogical roles and supports students’ intellectual responsibility. These findings have implications for policy development, particularly in child-friendly explainability, strict data governance, and effective human-in-the-loop procedures.