Artificial Intelligence in the EU: A Double-Edged Sword on the Path to Defence Autonomy

The European Union is facing increasing geopolitical pressure to strengthen its defence capabilities in a context where artificial intelligence (‘AI’) is rapidly transforming modern warfare. This creates a dual challenge: enhancing military AI capabilities while preserving the EU’s human-centric approach to AI. Although military AI has been identified as a priority, the EU lacks a unified position, and fragmentation continues to characterise its defence landscape. Institutionally, approaches diverge. The European Parliament (‘EP’) promotes a rights-based framework centred on human control and ethical constraints, while the European Commission (‘EC’) and the European Council emphasise capability development and defence readiness. Existing strategic frameworks and cooperation mechanisms do little to reconcile these perspectives. At the same time, the exclusion of military AI from the AI Act, the Union’s comprehensive AI legislation that regulates AI systems on the basis of the risk they present to fundamental rights, leaves governance largely in the hands of Member States, reinforcing regulatory and operational fragmentation. National approaches to military AI vary significantly, ranging from more developed strategies to indirect or absent references. In contrast, global actors such as the United States and China are advancing more coherent and centralised approaches, widening the gap with the EU. This uneven landscape risks not only capability fragmentation but also value fragmentation, as differing standards for responsible military AI could potentially emerge across the Union. This policy paper argues that current geopolitical momentum, increased defence spending, and the EC’s growing role in EU defence matters provide an opportunity to address these challenges. The EU should leverage its supranational strengths, particularly in funding and coordination, to foster a more coherent approach to military AI, including through a better understanding of its dual-use nature and the role of civilian AI actors. At the same time, financial and procurement instruments should be used to embed responsible, human-centric principles into military AI development. Linking EU support to ethical requirements would help align capability development with the Union’s values, addressing both defence and value fragmentation without requiring immediate legislative reform.