Abstract
This article examines artificial intelligence governance approaches adopted by the European Union, South Korea, and Japan, and evaluates their suitability for informing an AI governance framework for Thailand as a developing economy. Applying institutional economics theory—drawing on North’s framework of formal and informal institutional constraints, Coase’s analysis of transaction costs, and Williamson’s governance structures—alongside policy transfer theory and the concept of legal irritants, the article assesses three regulatory philosophies: the EU’s comprehensive risk-based hard law (Regulation (EU) 2024/1689), South Korea’s framework legislation integrating governance with industrial strategy (AI Basic Act), and Japan’s innovation-supportive promotional legislation (AI Promotion Act). Using comparative methodology structured around five dimensions—regulatory form and legal certainty, rights protection, innovation compatibility, institutional feasibility, and transferability—the article argues that Thailand should develop a context-sensitive hybrid framework through a phased approach combining voluntary principles, targeted binding obligations for high-risk AI, and a consolidated framework law. The article contributes to the literature on AI governance in developing economies by demonstrating how institutional economics analysis, combined with comparative regulatory assessment, can inform governance design balancing rights protection with innovation, institutional realism with normative ambition, and international alignment with domestic fit.