Best's Review


Book Store
Authors: Machine Learning, AI Offer New Opportunities in Risk Management

Risk professionals turn to machine learning and artificial intelligence to forecast the financial risks of extreme events.
  • Lori Chordas
  • November 2022

Risk Modeling: Practical Applications of Artificial Intelligence, Machine Learning, and Deep Learning

Terisa Roberts

Terisa Roberts

Stephen Tonna

Stephen Tonna

Technologies driven by machine learning and AI have transformed industries and help evaluate and solve risk management problems. Now risk managers and business leaders are also relying on those tools to evaluate the financial impact of extreme events such as pandemics and changes in climate, said Terisa Roberts, global solution lead-risk modeling and decisioning, and Stephen Tonna, senior risk consultant and risk lead, at SAS. They are also authors of a new book, Risk Modeling: Practical Applications of Artificial Intelligence, Machine Learning, and Deep Learning.

Following is an edited transcript of an interview with AM Best TV.

How did the idea for the book come about?

Roberts: While we see the use of AI and machine learning pervasive in many aspects of our society, at the same time the digital revolution is changing model paradigms and how we manage risk. Compared to the rapid adoption and growth of AI and machine learning in other areas of business, its uptake in financial risk management seems a bit slower. Risk practitioners are more cautious about its pertinence in regulated areas that also require transparency in decision making and are subject to high regulatory scrutiny. The aim of our book is to demystify the black box and highlight areas where it adds tangible benefit to the risk management function.

How can machine learning and AI be used to forecast and evaluate the financial risks and impact of extreme events?

Roberts: During the COVID-19 pandemic, machine learning proved very useful for short-term, high-frequency forecasting. These techniques can provide accuracy improvements and machine learning is very good at approximating complex risk calculations that might take hours to compute.

Tonna: They can also be used to supply faster, easier and more accurate insurance claims outcomes by using models to identify key questions needed to improve or deny medical coverage. They are effective at modeling complex and granular relationships at scale. That is what is needed in terms of our quantitative climate impacts on financial organizations' balance sheets. It is also important that risk models now make it to market in less than 12 months to help financial organizations respond to the market changes like those created by a pandemic and changing customer and regulatory expectations.

Lori Chordas is a senior associate editor. She can be reached at

AM Best Business Trilogy

AM Best details the history of AM Best, credit rating agencies and the life of Alfred M. Best.

The Company—A History of AM Best

The Industry—A History of Credit Rating Agencies

The Man—A Biography of Alfred M. Best

Go to Amazon to find these and other books.

Send us your book recommendations at

There’s So Much to Cover—Don’t Miss the Latest

Get more news stories like this delivered to your inbox by signing up for our article spotlights.


Back to Home