The emergence of Artificial Intelligence [AI] has transformed learning assessment practices, particularly in response to the need for authentic and differentiated evaluation in the digital era. While AI promises greater efficiency, its adoption in schools also raises several ethical and technical concerns. This article argues that AI will be most effective when positioned as a teacher’s partner through a Human-in-the-Loop approach, rather than as a fully automated replacement. To examine this issue, the study employs a literature review method with thematic analysis of 30 articles from SINTA and Scopus-indexed journals published between 2020 and 2025. The findings indicate that AI offers three significant opportunities: accelerating the assessment process through Automated Essay Scoring, adjusting item difficulty adaptively, and identifying at-risk students using predictive analytics. However, implementation faces three major challenges, namely cultural bias in algorithms, the risk of student data breaches, and the erosion of teachers’ sensitivity in assessment. Therefore, this article recommends clear regulations, strengthening teachers’ AI literacy, and a collaborative model between AI and teachers to ensure that assessment remains both accurate and humane.
Copyrights © 2026