Skip to content

Content Library

Back to Articles

Article July 8, 2025

AI Won't Save You... But It Might Spot Trouble Before You Do

AI Won't Save You... But It Might Spot Trouble Before You Do

Shameela Miah, Conference Producer

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

display-middle-ads-here

At CeFPro’s 2025 Risk Americas conference, amid the buzz around artificial intelligence and machine learning, one speaker managed to strip the hype down to its most pragmatic core.

Riten Dixit, Head of Market and Credit Risk Analytics at the Federal Home Loan Bank of Cincinnati, challenged the room to think not about how AI might revolutionize financial modeling, but how it could quietly – but powerfully – refine it.

“AI isn’t here to replace our traditional risk models – it’s here to augment them,” Dixit explained. For a room filled with modelers, validators, and risk strategists, this wasn’t a declaration of disruption but one of realism.

At a time when so many financial services organizations are looking for ways to implement automated solutions to meet myriad challenges, Dixit cautioned against viewing – and treating – AI as a silver bullet.

Instead, he advocated for using it as a supplement to conventional approaches, particularly where traditional models fall short.

He explained that while traditional risk modeling values parsimony – few variables, clean interpretability – AI thrives on complexity. It detects subtle shifts early, long before they show up in ratios or backward-looking indicators.

In sectors like mortgage finance, Dixit pointed to the erratic refinancing behavior during the pandemic and the divided borrower pool of today – some locked into low rates, others priced out entirely.

These, he argued, are structural anomalies that evade conventional modeling logic. Here, machine learning, especially anomaly detection techniques, delivered genuine value by flagging patterns too faint or nonlinear for legacy tools to catch.

Dixit also revisited the 2023 regional banking upheaval. As deposits shifted and balance sheets responded to rate shocks, AI tools proved prescient. “The real value of AI for us,” he said, “was the ability to pick up patterns on the balance sheet before they showed up in the financial statements.”

Sign Up to Our Newsletter

Sign up

Connect Plus

Unlock more with Connect Plus membership

This content is available to Connect Plus members.