Managing AI models is reshaping three lines of defense, say banks

The job of managing artificial intelligence models is challenging the traditional three lines of defense framework, say risk managers, with some calling for the first line to take on more formal responsibility for continuous testing and monitoring of AI models.

With financial risk models, traditionally, a bank’s model validation team creates a challenger model, and in some cases also does part of the monitoring and testing. But Commerzbank wants to “push a bit more of that back to the first line