This paper explores political drivers and policy process of the reform of the framework for Artificial Intelligence regulation and governance in the European Union (EU). Since 2017, the EU has been developing an integrated policy to tighten control and to ensure consumer protection and fundamental rights. This policy reform is theoretically interesting, raising the question of which conceptual approaches better explain it, and it is also empirically relevant, addressing the link between risk regulation and digital market integration in Europe. This paper explores the policy reform mainly by using two case study methods—process tracing and congruence procedure—using a variety of primary and secondary sources. It evaluates the analytical leverage of three theoretical frameworks and a set of derived testable hypotheses concerning the co-evolution of global economic competition, institutional structure, and policy preferences of domestic actors in shaping incremental approach to AI regulation in the EU. It is argued that all three are key drivers shaping the reform and explain the various stages of the policymaking process, namely problem definition, agenda-setting, and decision-making, as well as the main features of the outcome.

本文探讨了欧盟 (EU) 人工智能监管和治理框架改革的政治驱动力和政策过程。自 2017 年以来，欧盟一直在制定一项综合政策，以加强控制并确保消费者保护和基本权利。这项政策改革在理论上很有趣，提出了哪些概念方法可以更好地解释它的问题，并且在经验上也具有相关性，解决了欧洲风险监管和数字市场一体化之间的联系。本文主要通过两种案例研究方法——过程追踪和一致性程序——利用各种一手和二手资料来探索政策改革。它评估了三个理论框架和一组衍生的可检验假设的分析杠杆，这些假设涉及全球经济竞争、制度结构和国内参与者在塑造欧盟人工智能监管增量方法方面的共同演变。有人认为，这三者都是塑造改革的关键驱动力，并解释了决策过程的各个阶段，即问题定义、议程设置和决策制定，以及结果的主要特征。