This is an English-language technical report of a study reported in full, in Korean, as Research Project No. 2025-1-08 under Korea's Educational Development Special Zone — Digital Education Innovation Project. It introduces no data, analysis, or claim absent from the Korean report. The study asked whether AI-based adaptive learning requires a structured, machine-interpretable representation of the national curriculum, what such a representation would contain, and what it would cost. It reports a needs analysis with in-service Korean teachers (a survey of 26 teachers, focus group interviews, and LDA topic modelling of the interview transcripts), an ontology schema prototype for the 2022 Revised Korean National Curriculum, a cost model for construction and maintenance across school subjects, and a working agent prototype that queries the ontology at generation time. Teachers rated factual error and misalignment with curriculum standards as the most serious problems with generative AI in educational use, and rated curriculum-aligned content generation as the most important function an ontology-based agent could provide. Exploratory factor analysis of the problem items returned three factors: alignment and personalization, reliability, and data and system control. Initial construction of curriculum ontologies for all 24 Korean middle-school subjects is estimated at approximately KRW 3.94 billion, with annual maintenance at approximately KRW 409 million. The study establishes that teachers report the need and specifies what an implementation would involve. It does not establish that an ontology-grounded system improves learning or assessment outcomes; no outcome was measured. Funded by the Educational Development Special Zone — Digital Education Innovation Project, Ministry of Education and Daegu Metropolitan Office of Education, and the AI–Digital Convergence Education Innovation Platform, Kyungpook National University. Project No. 2025-1-08.
Jaehwa Choi, Kyungil Yoon, Hong-Gee Kim et al.· Zenodo (CERN European Organi...· 4 citations
This is an English-language technical report of a study reported in full, in Korean, as Research Project No. 2025-1-09 under Korea's Educational Development Special Zone — Digital Education Innovation Project. It introduces no data, analysis, or claim absent from the Korean report. The study compared the centralized platform (CP) model, as instantiated in Korea's AI Digital Textbook programme, with a decentralized ecosystem (DE) model in which teachers design, build and operate their own generative AI agents. Work ran from June to November 2025 in four stages: theoretical grounding; a survey and focus group study of teacher perception and requirements; a teacher-led agent development programme; and the formulation of a CP–DE hybrid framework. Twenty-four teachers completed the first survey; 22 had used the AI Digital Textbook. They rated the difficulty of getting teacher feedback reflected in the platform as its most serious limitation (79.2%), while recognising access (45.8%) and curricular consistency (41.7%) as genuine strengths. Perceived necessity of a teacher-led alternative was M = 4.71 and intention to participate M = 4.67. Topic modelling of focus group transcripts (K = 5) returned topics dominated by the limits of centralized AI and by teachers' identification with a maker rather than user role. A 30-session capacity-building track was developed and run, and participating teachers designed and deployed working agents in their own classrooms. The study establishes what teachers report needing and demonstrates that teachers without programming backgrounds can build usable agents. It does not measure whether teacher-built agents improve assessment quality or learning outcomes, and it contains no operational record of how often teachers revised or rejected agent output. Funded by the Educational Development Special Zone — Digital Education Innovation Project, Ministry of Education and Daegu Metropolitan Office of Education, and the AI–Digital Convergence Education Innovation Platform, Kyungpook National University. Project No. 2025-1-09.
Jaehwa Choi, Kyungil Yoon, Sung‐Yeon Kim et al.· Zenodo (CERN European Organi...· 4 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
