The European Central Bank (ECB) has updated its Guide to Internal Models to reflect new priorities in banking supervision. A key change is the stronger focus on ESG risks, especially climate and environmental factors. Banks are now expected to check whether these risks are material at every stage of their internal models and, where relevant, include them in how they calculate capital requirements. This means the ECB wants banks to treat climate and environmental risks as seriously as traditional financial risks when assessing creditworthiness and resilience.

The guide also sets out clear expectations for the use of machine learning (ML) in internal models. While ML offers powerful tools, the ECB warns of risks like lack of transparency, model drift, and overfitting. To address this, banks must strengthen governance and validation, ensure management and oversight teams understand ML, and apply explainability tools so that complex models remain reliable and accountable. In short, the ECB is signalling that innovation is welcome—but only with strong safeguards