ECB and ESRB warn banks about AI-powered cyber attacks

In short: Europe’s main banking watchdogs say advanced AI could help hackers find and use weak spots in bank systems within minutes, and they want banks to respond fast.

What happened

Europe’s top bank supervisors, the European Central Bank (ECB) and the European Systemic Risk Board (ESRB), issued new warnings about cyber attacks that use “frontier” AI models. Frontier models are the most capable general AI systems available today.

Claudia Buch, who leads the ECB’s banking supervision board, sent a letter to 110 banks. She gave them until the end of October to produce a “comprehensive action plan” to deal with these risks. The letter asks for specific steps, enough staff and budget, clear ownership inside the bank, and deadlines for completing the work.

In a separate warning, the ESRB said newer AI models can find and exploit software bugs much faster than older tools. A bug is a mistake in software (like a broken lock on a door). The ESRB warned that these weaknesses could now be “weaponised” in “minutes or hours,” which could threaten financial stability across the Eurozone.

Buch also told banks to be ready to detect and stop large-scale attacks quickly, and to make sure outside tech suppliers can do the same. She said banks should also improve “cyber hygiene,” meaning everyday basics like keeping systems updated, limiting access, and monitoring for suspicious activity.

To help with the tight timeline, the ECB is delaying its annual IT risk questionnaire. Banks now have until February to submit it, instead of September.

Why it matters

Banks are a central part of daily life, from paychecks to card payments. If AI helps attackers move faster, problems that used to be caught in time could spread more quickly, which can lead to outages and disrupted services for customers.

Source: Financial Times