BETWEEN HARMONISATION AND AUTONOMY:

Abstract

Artificial intelligence (AI) has emerged as a transformative force in public administration. While an “AI spring” promises efficiency, integrating these systems into governance ignites global competition and raises critical questions about accountability and democratic control. Existing research has focused heavily on technical implementation and maturity indices, frequently overlooking the discursive processes that define the very “problems” AI is expected to address. However, a significant research gap exists regarding how a state’s institutional relationship to the European Union shapes these representations. Specifically, the discursive divergence between EU membership and post-Brexit regulatory autonomy remains under-researched within the field of European Studies. By integrating governance and framing theories, this thesis addresses this gap by applying Carol Bacchi’s “What’s the Problem Represented to be?” (WPR) approach to a comparative case study of Swedish and British policy documents (2018-2025). The findings suggest that while Sweden represents AI as an existential welfare necessity embedded in the EU’s “Russian Doll” of harmonisation, the United Kingdom represents it as a lever for market-led sovereignty through a “sovereign illusion” of post-Brexit agility. These differing paths appear to constitute distinct subjectivities: positioning the Swedish citizen as a passive data provider and the British citizen as a market consumer. This study demonstrates that AI governance is a discursive tool used to negotiate state legitimacy within the European regulatory landscape. By framing AI as a neutral infrastructure or an unavoidable geopolitical race, both states effectively depoliticise digital governance, moving accountability from the democratic sphere to technocratic or commercial domains.