Preparation for Introducing Generative AI into the Next-Generation Educational App “A-txt” for Smart Devices

This study explores the integration of generative AI (large language models) into the AR-based educational support application “A-txt” to develop a next-generation personalized learning support system utilizing learners’ behavioral and question histories. A-txt recognizes charts and texts on printed educational materials as AR markers and displays associated digital content. In this paper, we present a prototype system in which generative AI provides real-time hints and explanations based on user interactions and natural language queries. Through experiments involving image recognition and learning support responses, we verify the effectiveness of the system. We also discuss future developments, including automatic content recommendations, visualization of weak areas, and feedback functions for educators based on learner behavior logs. This research provides a foundational step toward building a context-responsive and self-regulated learning support environment by integrating AR and generative AI technologies.