Cahaya Nusantara Primary School, Bandung, Indonesia

Abstract: Against the policy backdrop of Indonesia’s Digital Education Roadmap, this study constructs a three-dimensional collaborative AI teaching model covering localized context design, technical adaptability and cultural inclusiveness, targeting core teaching pain points including difficulties in abstract mathematical concept comprehension and unequal access to educational resources between urban and rural areas. Based on two empirical cases — fraction teaching in public primary schools of Jakarta and geometric teaching in rural schools of Yogyakarta, this paper systematically explores the innovative applications of generative AI in reconstructing localized teaching content, adapting to limited technical environments and integrating indigenous cultural values. The experimental results demonstrate that the proposed model can significantly boost students’ mathematics learning motivation (32.6% growth in the experimental group) and conceptual understanding (27.4% higher test scores), with more prominent improvements in learning engagement among rural students (an increase of 38.2%). This research puts forward an AI teaching maturity ladder tailored to Indonesia’s education system, providing progressive practical paths for frontline teachers ranging from basic tool operation to in-depth instructional design. It also confirms that the cultivation of teachers’ digital literacy and the development of localized teaching resources are the core prerequisites for sustainable AI application in education.

Keywords: artificial intelligence; primary mathematics; digital teaching; educational technology; localized teaching