When National Security Becomes a Shield for Evading AI Accountability

As AI is integrated into state security across Europe, the existing legal protections are not keeping pace, leading to significant regulatory gaps.
EU member states increasingly leverage AI for automated decision-making, expanded surveillance, and consolidating state power without adequate regulation.
Existing EU legislation includes broad exemptions related to “national security” which states exploit to bypass oversight on AI usage in surveillance and counterterrorism.
The Court of Justice of the European Union (CJEU) provides guidance on national security but limits its use to genuine and present threats.
Specific cases in France, Hungary, Serbia, and Spain exemplify how national security is invoked to justify AI deployment and surveilential measures, often with minimal public oversight.
Landmark CJEU cases like La Quadrature du Net and SpaceNet set strict conditions for AI systems used for national security purposes.
To ensure accountability, civil society should uncover details of AI-based national security initiatives, engage in strategic litigation to enforce existing regulatory standards, and build cross-border coalitions for shared advocacy strategies.
Closing the loophole in the EU AI Act that allows governments to bypass AI oversight by classifying systems as “national security” tools is essential for protecting fundamental rights and democratic oversight.