The European Central Bank (ECB) has published a special feature assessing how advances in artificial intelligence (AI) could overwhelm cyber defences and amplify systemic risks for euro area banks. AI agents can increase the speed, scale and sophistication of attacks while creating vulnerabilities within banks’ own systems. Although AI also improves vulnerability detection and patching, costs, governance challenges and slower deployment decisions may leave defenders at a structural disadvantage in the near term.
The euro area banking system has not experienced a systemic cyber event, but unequal access to advanced AI models and differences in banks’ defensive resources create vulnerabilities. Reliance on a few technology providers, predominantly outside the European Union, adds potential single points of failure. Around a fifth of cyber incidents affecting euro area banks originate from external providers. AI could make simultaneous attacks on multiple institutions easier and amplify their effects through misinformation that undermines customer or counterparty confidence.
The feature calls for stronger controls over AI agents, closer monitoring of interactions with AI models and defences that assume systems may already be compromised. It also recommends coordinated contingency planning across banks, technology providers and authorities, alongside frequent testing and cyber mapping to identify common exposures in technology supply chains. These priorities echo ECB President Christine Lagarde’s call at the European Systemic Risk Board’s October conference for updated cyber defences and coordinated response plans.