Seocho High School, Seocho-gu, Seoul, Republic of Korea

Abstract: Despite South Korea’s consistently high rankings in international assessments such as TIMSS and PISA, high school physics education faces a persistent “formula-centric dilemma,” where students demonstrate strong procedural fluency but lack deep conceptual understanding and real-world application skills. This paper introduces the “AI-Empowered Three-Stage Modeling Cycle” (Predict-Model-Verify), an innovative instructional framework designed to bridge this gap. By integrating artificial intelligence tools, such as PhET interactive simulations and AI-driven data analysis assistants, with culturally and technologically relevant Korean contexts, this model transforms passive learning into active scientific inquiry. Two specific instructional cases are detailed: the application of projectile motion to Gungdo (traditional Korean archery) and the analysis of fluid dynamics using the KTX-Sancheon high-speed train. Empirical results from a quasi-experimental study (N=80) indicate that students in the experimental group achieved a 25% greater improvement in scientific inquiry skills and a significantly higher normalized gain on the Force Concept Inventory (FCI) compared to the control group. This paper outlines the theoretical foundations, instructional design, empirical outcomes, and critical reflections on the integration of AI and local context in Korean physics education, offering a scalable model for future curriculum reforms.

Keywords：AI-Empowered Instruction，Physics Education，Modeling-Based Learning，2022 Revised Curriculum，Gungdo (Traditional Korean Archery)，KTX High-Speed Train，Scientific Inquiry Skills，Process-Oriented Evaluation，Culturally Relevant Pedagogy，Digital Textbook Initiative