Data Science & AI

Listen: MIT Tech Review's Podcast Looks At How AI Can Fix Credit

Zest AI team
May 12, 2021

Zest AI CEO Mike de Vere was featured today on MIT Technology Review's award-winning podcast In Machines We Trust, which looks at all the ways algorithms and automated decisions are re-shaping society and our economy. The latest episode explores the risks and challenges posed by credit scoring algorithms and their use in decisions ranging well beyond finance and lending. But, if built thoughtfully and with an eye toward fairness, AI/ML algorithms can make a dramatic impact on financial inclusion.

The hosts of the show spoke with Zest about its work reducing the disparate impact that's baked into historical credit data using a technique called adversarial de-biasing. We're helping lenders build more inclusive models that have safely increased approval rates for women and applicants of color by more than 25%. We're also develeping new algorithms that do a better job than legacy techniques at identifying the race of borrowers when race data is not allowed to be collected in loan applications (i.e., any loan that's not a mortgage). In one test with Florida's voter database, Zest was able to improve the accuracy of race prediction by 60%. This will have a huge impact on fair lending in the future.

Give a listen and let us know what you think. You can subscribe to the show here.

September 12, 2023
Why picking the right AI-credit decisioning partner matters
September 7, 2023
Machine learning 101: Using ML to increase portfolio yield
September 6, 2023
Credit Unions
Tech costs should not be killing innovation for small credit unions