AI Won't Save You... But It Might Spot Trouble Before You Do
At Risk Americas 2025, Riten Dixit of the Federal Home Loan Bank of Cincinnati laid out a clear-eyed vision for how AI and machine learning can enhance—not replace—risk modeling. His message? In volatile markets, AI’s real value is early detection and bet
- AI and ML are being used to augment, not replace, traditional risk models
- Riten Dixit
emphasized AI’s strength in identifying early signals in complex markets
- Traditional
models fail to capture structural shifts like today’s mortgage market anomalies
- During the
2023 banking crisis, AI spotted risks before they appeared in financials
- Validation
of AI models should focus on decision usefulness, not just accuracy
- Dixit
developed a clustering tool using NLP and learning algorithms for identifying
bank vulnerabilities
- AI enhances
pattern recognition in fragmented and fast-changing risk landscapes
- Future
modelers must act as architects, using diverse tools to build decision-ready
frameworks
- Emphasis was
placed on communication and transparency with AI model users
- The real
value lies in models that improve thinking, not in achieving technical
perfection
Sign in to continue reading
What’s included
- Access to peer-contribution articles and insights
- Access to the latest global research and market intelligence reports
- Access to the latest Connect Magazine, a monthly publication
- Insight articles, panel discussions, webinars, podcasts and peer-led interviews
Connect Plus Membership - $199
Become a Connect Plus member to unlock more insights from our knowledge hub, receive 15% discount on all events and access to audience polls and speaker presentations for up to three CeFPro events.