ECB Rethinks AI Assurance in Banking Amid Accelerating Adoption

The European Central Bank (ECB) is rethinking its approach to artificial intelligence (AI) assurance in banking, citing that traditional model validation methods are not sufficient for generative and agentic AI systems.

According to Pedro Machado, the ECB's representative to the Supervisory Board of the Single Supervisory Mechanism, over 85% of large banks under European supervision already use AI in some form. The adoption is accelerating as these systems move beyond experimentation.

The ECB warns that existing governance and model risk frameworks cannot be simply expanded without accounting for the distinct behavior, dependencies, and lifecycle risks introduced by generative and agentic AI.