Exploring Design Principles for Generative AI-Supported Reading Education: An Illustrative Lesson Design

Abstract

This study aims to analyze recurring design issues in reading education using Generative Artificial Intelligence (GAI) and to structure them into design principles applicable to reading instruction. To this end, 47 core studies were analyzed, and meaning units, initial codes, subcategories, and provisional upper categories were derived. Subsequently, a first-round independent rating and a second-round re-rating were conducted with six experts in reading education, Korean language education, AI literacy, AI ethics, educational technology, and educational evaluation. In the second-round re-rating, all 40 items showed an I-CVI of .83 or higher, and the S-CVI/Ave was .96. As a result, five design principles for GAI-based reading education were derived: learning support and performance efficiency (Efficiency type), creative expansion of reading experience (Augmentation type), critical verification of AI-generated output (Critique type), protection of learner-directed reading (Detachment type), and adaptive learning support (Personalization type). These principles were also explained as being combined through adaptive regulation, a complementary pathway, auxiliary interactions, and tension– balance relationships. Based on these findings, a one-period instructional design example was presented for argumentative text reading in the second year of middle school. This study is significant in that it specified the core of GAI-based reading education not as the introduction of AI functions but as the design of boundaries between AI support and learners’ direct performance.