Fair Lending Regulations with AI in Banking

Banking CIO Outlook | Tuesday, July 30, 2019

With AI and ML entering many aspects of the lending ecosystem, bankers are transforming their legacy lending approaches to drive success through intelligent, fair-lending principles.

FERMONT, CA: The next wave of opportunities in the global finance industry is AI. In many stages of the lending arena, AI plays an essential role from loan approval and asset management to risk assessment. This article explains some of the ways the consumer lending industry is taking this technology by storm.

One instance of how economic regulators ensure fairness by defending against discriminatory loans is the legitimate prohibition of ZIP codes in underwriting. But the growing dependence on AI and machine learning, or "automated insights," has made testing a broad range of inputs for particular results a scientific process by businesses.

Cash flow information may remove the banker's unique features, affinity, and attitudes. Instead, it makes a decision based on the company's past and present results, as shown by dozens of holistic touchpoints in the information. These models are still extremely controlled by humans, but the algorithms can use non-traditional or alternative data. As clients now produce a sea of precious data, more holistic information can be achieved through nuanced and objective choices. Online lending platforms are very cautious with the data used to identify the most critical lending aspects: the capacity of a borrower to repay continuously and independently, a readiness to repay rather than default.

Companies test to discover variety or consistency through various trials. Data researchers are diligently reviewing the results of unnecessary bias indices in real-time, while regulators are waiting for outcome patterns that may arise for years. Most data-driven fintechs already provide inputs and responsibly controlled results with a culture of testing, modeling, and principles-based examination. Regulations arbitrarily restricting data inputs only hold back progress, especially for the billions of underbanked adolescents attempting to access official financial services.  

If the results fulfil legal requirements, reasonable use of that information improves access to robust banking facilities for small companies, rather than excluding or penalizing them.

Weekly Brief

Read Also