Europe’s top banking watchdog just told banks they need a plan for frontier AI. Not a “let’s circle back on this” kind of plan. A real one, with contingencies, cybersecurity upgrades, and operational resilience baked in.
The European Banking Authority published its Spring 2026 Risk Assessment Report on June 18, flagging frontier AI large language models as a growing threat to the financial system. The core concern: these models are getting good enough to find and exploit software vulnerabilities faster than human security teams can patch them.
What the EBA actually found
The EBA is urging banks to strengthen three pillars: operational resilience, cybersecurity measures, and contingency planning.
The European Central Bank raised similar concerns in May 2026, going so far as to initiate direct conversations with major banks about AI-driven cyber risks. The ECB’s message was blunt: frontier AI capabilities could outpace even skilled cybersecurity practitioners.
