The Black Box Issue with AI is Starting to be Solved

One of the biggest impediments to using AI to lend is that credit decisions are made in a black box and they cannot be explained; a number of fintech firms and big banks are helping to change this and offer explainability so regulators can get comfortable with AI driven underwriting; regulators want to understand if someone is denied credit then what is the reason for denial; the fintech firms argue that a lot of the current metrics like income and FICO score are already skewed towards to most well off people; ZestFinance, Upstart and Underwrite.ai use a variety of different data points in their decision engines and all offer explainability as to why someone is denied credit; AI has started to make a turn towards wide scale adoption but it still has a ways to go and regulators need to ensure they keep up with this change. Source.

  • Todd Anderson

    Todd is the host of PitchIt: the fintech startups podcast, a weekly interview show featuring emerging fintech founders and leading venture capitalists. He is responsible for leading the content team which covers fintech through daily & weekly email newsletters, editorial, virtual events, and in-person conferences. He has been covering fintech, banking, and venture capital for more than 15 years, including speaking regularly at industry events.