Meridian Grace Primary School, Singapore

Abstract: With the advancement of Singapore’s Smart Nation 2025 education strategy, traditional mathematics assessment models are confronted with prominent challenges including insufficient timeliness and coarse diagnostic granularity. Based on the dynamic analytical capability of generative AI and the motivational stimulation mechanism of gamified contexts, this study constructs a three-dimensional assessment model featuring “diagnosis-intervention-prediction”. A controlled experiment conducted in three primary schools in Singapore (N=240) verified that the proposed model increases students’ problem-solving transfer rate by 37% and accelerates the development of computational thinking by 28%. The core innovations of the study are threefold: first, developing a generative AI-powered personalized error cause mapping technology; second, establishing a cross-culturally adaptive framework for gamified assessment; third, realizing the real-time transformation from assessment data to instructional decision-making.

Keywords: Generative AI; Gamified Assessment; Error Cause Mapping; Dynamic Measurement; Singapore Mathematics Education