AI Governance Demands A New Approach To Model Risk Management

In this onsite interview from Risk Americas, Riten Dixit sat down with Alana Cannatella, Marketing Manager at CeFPro to explore and examine the growing challenges AI presents for model risk management. He explains why traditional model inventories are becoming increasingly difficult to maintain as organizations deploy AI models at greater scale and speed, and highlights the importance of identifying dependencies, assessing materiality, and understanding the potential impact—or "blast radius"—of individual models.

Dixit also discusses the future of AI governance, arguing that organizations must move beyond periodic validation cycles toward more continuous, risk-based oversight. He explores how AI introduces risks across multiple enterprise risk categories, including market, credit, cyber, operational, and reputational risk, and why firms need more transparent and scalable governance frameworks to manage these emerging challenges effectively.