RAG-ENHANCED SOCRATIC ADAPTIVE CHATBOT FOR HISTORICAL INQUIRY LEARNING WITHIN LEARNING MANAGEMENT SYSTEM

Abstract

The development team comprehensively designed the Patih AI platform to revolutionize Phase F Indonesian History learning at the high school level. This digital system integrates adaptive artificial intelligence technology to holistically address the problem of rote memorization. Preliminary surveys show that 92.9 percent of 11th grade students have difficulty learning due to the low relevance of everyday material. Researchers combined the Successive Approximation Model (SAM) approach and design thinking mechanisms during the development process of this educational software system. The Retrieval-Augmented Generation (RAG) architecture supports the Socratic mentor function to guide students' reasoning processes in virtual interaction features. Kinesthetic gamification features strengthen long-term memory retention capacity through instant detection of students' body movements. The research team tested the feasibility of this system on 312 students from five schools in East Java. Field evaluation results recorded a 35.03-point increase in historical critical thinking scores with a statistical significance value of 0.005. Testing of the system's usability instrument resulted in a feasibility score of 74 for operational readiness assurance at the large-scale implementation stage. This innovative platform facilitates a meaningful history learning experience without creating the risk of cognitive dependence for users.