AI Textbooks and a Textbook Case of Policy Failure in South Korea

• In Event: AI and Educational Innovation Ecosystems

Abstract

South Korea’s AI textbook initiative, launched as a flagship project in digital education reform, has encountered serious implementation challenges. Despite strong governmental backing, many schools declined to adopt the materials, citing poor usability, weak curricular alignment, and limited pedagogical value. Teachers questioned the functionality of the AI tools, and legal disputes between IT firms and the Ministry of Education revealed deeper structural misalignments. The initiative’s eventual collapse also triggered significant layoffs in the edtech sector. Against this backdrop, the study asks: What mechanisms contributed to the failure of South Korea’s AI textbook initiative to stabilize as a sustainable educational innovation?

To answer this question, the study first reviews existing research on digital technology adoption and policy failure, identifying a disconnect between prevailing frameworks and the complex dynamics observed in practice. To address this gap, it develops a theoretical framework that draws on megaproject theory, technological innovation theory, and policy symbolism to analyze how political ambition, sociotechnical breakdown, and organizational misalignment interact. This framework informs a hybrid methodology combining process tracing and Actor-Network Theory (ANT), which together reconstruct both institutional trajectories and the role of non-human actors in the network of failure.

The study finds that policy failure emerged through four intersecting dynamics. First, the government overestimated the promise of innovation without sufficient planning, reflecting a lack of institutional readiness for digital transformation. Second, symbolic goals—such as political ambition and international image-building—contributed to distorting the policy’s priorities and overshadowing its pedagogical aims. Third, non-human actors—such as AI textbooks, digital platforms, and legacy systems—disrupted implementation by failing to integrate or function as intended. Fourth, the absence of institutional strategies for managing partnerships with lean-agile IT firms revealed a structural blind spot in public innovation governance.

By tracing these dynamics across institutional and sociotechnical domains, the study contributes to debates on digital governance, symbolic reform, and the infrastructural conditions under which innovation fails.

Author

• Youngdal Son, Sungkyunkwan University
  Presenting Author