Model Risk Management: Today's Governance and Future Directions

The applications of artificial intelligence, machine learning and automation in risk and compliance functions are pervasive. In order to deal with the rising number of models, model risk management leaders are often challenged to adapt to the changes required in model policy, appetite and culture. How can firms build trust and confidence in their models based on the right data, models and controls at the enterprise level? This is especially important for high materiality models driving business outcomes. The SAS/GARP AI/ML Model Risk Management survey investigated where firms are heading. It is clear that the introduction of new techniques, such as AI/ML, and the frequency of model updates (from months to days) will quickly outpace the current fragmented model governance infrastructures. To keep up and have a stronger view of model risk, a fundamental infrastructure reboot is necessary.