When the EU AI Act Travels: Legal Transplantability and Modular Adaptation in Third-Country AI Governance

Abstract

This article examines the extent to which the EU AI Act can function as a model for AI governance beyond the European Union. It argues that regulatory influence should not be equated with successful legal transplantation. While the Act has quickly become a reference point for lawmakers and regulators outside Europe, its elements vary significantly in portability across legal systems. The article develops and applies a four-part framework for assessing legal transplantability, focusing on normative transplantability, institutional transplantability, market transplantability, and rights-operability. It uses selected third-country examples, including Japan, Singapore, the United Kingdom, and Brazil, not as an exhaustive survey, but as comparative tests of how different regulatory environments absorb, translate, or resist elements of the EU model. On that basis, the article distinguishes between elements of the Act that are comparatively more portable, such as risk-based classification, baseline transparency duties, documentation logic, and incremental governance tools, and those that resist transplantation more strongly, especially enforcement density, technical conformity infrastructures, and the rights-operability that makes fundamental-rights commitments actionable in practice. The article concludes that the EU AI Act travels most plausibly not as a complete statutory template, but as a modular regulatory architecture. Its comparative value for third countries lies less in wholesale replication than in selective adaptation sequenced to institutional capacity and regulatory readiness.