Part 1 of this paper first provides an update on the progress of AIDA and the EU AI Act (I), outlines a summary roadmap of the base similarities and differences between the two items of legislation (II), reviews the objectives of AIDA in contrast to the EU AI Act (III), compares their respective jurisdictional scope (IV), reviews their respective definitions of AI systems (V), outlines new definitions/concepts which have been introduced into the legislation (VI), outlines the extent to which there are exclusions for the public sector (VII) and for research (VIII), and provides an overview of their respective risk-based approaches (IX).

The aim of this article is to assess the current global reach of the European Union’s regulatory influence on AI. This analysis focuses on how an “AI system” is defined in the final EU AI Act and presents the hypothesis that the more rigid the definition of AI systems is, the less global the reach of the EU AI Act’s standards will be, especially in countries with a strong tendency towards AI sovereignty. A qualitative comparative analysis of four case studies (Brazil, Canada, Chile, USA) reveals persistent divergences centred on the material scope of AI systems. In turn, the global reach of the EU AI Act seems rather limited at best in the long term.

After a brief introduction (I.) and a closer look at the exact timeline (II.) of the Artificial Intelligence Act, the article addresses key questions ranging from regulated AI applications over the risk-based approach and the classification system to the wide range of high-risk AI systems (III.) before a conclusion concerning the likely impact of the AI Act is drawn (IV.).