Thursday, May 24, 2018

AI Challenging Bank Lending Practices

&l;p&g;&l;img class=&q;dam-image shutterstock size-large wp-image-1093777685&q; src=&q;https://specials-images.forbesimg.com/dam/imageserve/1093777685/960x0.jpg?fit=scale&q; data-height=&q;720&q; data-width=&q;960&q;&g; Shutterstock

Artificial Intelligence may be pushing the limits of decision making. Banks that want to use AI still have to adhere to several industry specific rules primarily around transparency. Other industries like tech don&a;rsquo;t have the same hurdles to climb when it comes to regulation around there AI efforts. Rules for banks were put in place to protect poor and minority consumers, including laws which are supposed to ensure the equal treatment of customers. Are black people going to fall further behind in getting approved for loans and credit cards if banks start using AI for lending decisions?

&l;b&g;The Breakdown You Need to Know&l;/b&g;

&l;span style=&q;font-weight: 400;&q;&g;Trying to interpret the mystery behind why deep learning decisions are made has proven to be quite difficult. AI already has a history of bias against minorities and if you add in how underbanked black communities are, putting this combination together could present major problems. There are laws like the Equal Credit Opportunity Act, &a;nbsp;which states banks have to prove they&a;rsquo;re not discriminating based on traits like gender or race. If banks let an AI system make a judgment about customer needs without being able to explain how it made the decision, inherent bias will control financial lending decisions. &a;ldquo;We want to understand how the decision is made, so that we can stand behind it and say that we&a;rsquo;re not disfavoring someone,&a;rdquo; said Hari Gopalkrishnan, client-facing platforms technology executive at Bank of America during its tech summit.&l;/span&g;

&l;a href=&q;https://www.culturebanx.com&q; target=&q;_blank&q;&g;CultureBanx&l;/a&g; notes AI platforms taking over bank lending decisions could put black family wealth further behind. The latest figures from the U.S. Census Bureau show the median net worth for an African American family is $9,000, compared with $132,000 for a white family. Latino families did not fare much better at $12,000. In 2016, 20.9% of black borrowers and 15.5% of hispanic borrowers were rejected for a loan, along with 8.1% of white applicants and 10.4% of Asian applicants, according to a report from Zillow. The gap is significant because owning a home can help most people, regardless of race, accumulate assets to boost their net worth.

What is deep learning technology? Let&a;rsquo;s start with the tools used for deep learning, which include neural software networks whose structure roughly tries to simulate the human brain&a;rsquo;s operations. These systems create outcomes with unprecedented accuracy and speed. However, the main issue here not always clear how the dense web of computations reaches a specific decision.

&l;b&g;AI Lending Mindfulness&l;/b&g;

&l;span style=&q;font-weight: 400;&q;&g;Banks in the past have argued they find themselves in a bind when staring down potentially conflicting government mandates to reduce balance sheet risk, while not discriminating against any one demographic group. As financial institutions dive further into deep learning technology, Bank of America and Capital One have AI researchers working on how to make the opaque technology transparent enough for wider use.&l;/span&g;&l;/p&g;

No comments:

Post a Comment