<![CDATA[Purpose – This study examines the integration of Artificial Intelligence (AI)-driven deep learning as both pedagogy and enabling technology within Indonesia’s middle school Informatics curriculum, responding to the shift from teacher-centered instruction toward adaptive, personalized learning.Methods – Using an embedded mixed-method design, the study combined a systematic literature review of 50 Scopus-indexed articles (2020–2025) with a 12-week quasi-experimental intervention at SMP Tahta Syajar, Bekasi. The proposed model comprised an AI-literacy-based learning framework, a Deep Neural Curriculum Loop (DNCL), and an adaptive feedback system. Quantitative data (ANCOVA) were triangulated with qualitative evidence from student portfolios and teacher interviews. Findings – The DNCL model significantly improved student engagement (+27%) and creative problem-solving (+31%) compared to conventional instruction, with large effect sizes for creativity (d = 1.5) and engagement (d = 0.8). AI-supported formative feedback enabled more authentic assessment of 21st-century skills and strengthened human–machine collaboration aligned with the Merdeka Belajar framework. Research Implications – The findings provide empirical support for AI-driven curriculum innovation in Indonesian secondary education; scalability is limited by teacher readiness, infrastructure variability, and the single-site research design.]]>
Copyrights © 2026
