Executive summary

Artificial intelligence and machine learning (AI/ML) are expanding the frontiers of finance. Over the next few years, we foresee a proliferation of AI/ML use cases in the back, middle and front office functions at banks.

Surveys show growing recognition among finance executives that adoption of AI/ML and use of advanced analytics will be the next big differentiator among competitors.

But amid the hype, banks need to tread with caution. For, AI/ML models come loaded with risk and complexity. The traditional or existing model risk management (MRM) frameworks are plainly insufficient to handle them.

This white paper uses the foundational building blocks of a traditional MRM framework as the point of departure to identify the precise areas where the framework must adapt to new AI/ML model risks across the model lifecycle.

It also recommends specific techniques where adjustments are due at each stage, in order to validate such models.