The EU–NATO regime complex and its influence on national AI strategies in Europe

Abstract

As artificial intelligence (AI) becomes increasingly crucial in global politics, states develop national strategies to define their approach to this technology. This development does not occur in a vacuum and is influenced by external normative frameworks from various international organisations (IOs), leading to regime complexity in which overlapping mandates and norms can complicate governance, particularly given AI's dual-use nature for both civilian and military applications. The EU–NATO regime complexity exemplifies this, as states with dual membership must navigate contrasting civilian EU norms and military NATO principles that address the same issue area – AI. The study aims to analyse how this regime complexity affects AI governance in the national strategies of four European states: Ireland (EU-only), the UK (NATO-only), and dual members Estonia and Croatia. Based on the concepts of regime complexity and institutional overlap, this study employs qualitative discourse analysis to analyse over 50 official documents covering the period 2017–2026. Analysis shows that all states implement the norms of the IOs they are members of and EU–NATO regime complexity affects the formulation of national AI strategies by leading states to adopt mixed AI strategies that include normative elements from both the EU and NATO. However, Croatia's case demonstrates that there are potential mediating factors that can influence this process, as evidenced by its current predominantly civilian approach despite the presence of military NATO elements in strategic documents. Notably, all AI governance begins with a civilian focus, suggesting that civilian discourse precedes defence one, which explains the UK's mixed strategy case. These insights contribute to regime complexity theory by introducing a technological dimension and by highlighting the need to further explore mediating factors.