The Politicization of the EU’s AI Act

So far, Artificial Intelligence remains shrouded in uncertainty. This applies to the technology’s potential capabilities, its societal and economic potential, as well as to the dangers it may create. This uncertainty also very much affects the ways that AI can be regulated and the societal concerns to which this regulation is to respond. When the European Union adopted its AI Act in 2024, it celebrated it as a ground-breaking act that would lay down a world-wide standard for AI regulation. However, the AI Act and its implementation very much remain work in progress. Also, it remains an open question whether the AI Act really reflects the concerns about AI that are prominent in the EU member states and whether it has been able to frame these concerns in ways that adequately respond to them. To identify and explain variation in the way the decision-making on the AI Act has been covered and politicised in different national public spheres, this paper examines the coverage of the legislative process of the AI Act in major broadsheets in four EU member states: Germany (SZ), France (Figaro), Netherlands (NRC) and Ireland (The Irish Times). Specifically, the analysis looks at the amount of attention that the AI Act has received, the concerns that have been brought to the fore, the ways in which conflict about AI is framed, and the main policy actors that are focused upon. The analysis then evaluates whether the variation in the politicization of the AI Act can be explained by the economic prominence of AI and AI-related industries in a country and/or by the presence of a national regulatory infrastructure for digital technologies. Furthermore, I examine the extent to which the political choices in AI regulation are framed as pitting a (presumed) collective EU interest against some external (technological or economic) opponent or rather as involving trade-offs among competing interests among EU citizens.