How AI creates fair lending

Financial inequality remains an unsolved issue. The lack of affordable housing has led to a crisis where long-term wealth generation through homeownership is impossible for many. Closing the wealth inequality gap, which is closely linked to race, has become increasingly difficult due to systemic biases within financial institutions. 

The bias exists because banks and mortgage companies reward simple and straightforward application processing. These rewards include discounted fees and lower rates. The rewards conversely discourage qualified yet complex mortgage loan applications that are more often submitted by minority group members with lower incomes.  

These issues are further aggravated by the processes and practices that are ingrained in mortgage and banking institutions. This was illustrated by the recent probe into mortgage bankers’ use of loan discounts or price exceptions which have not been equally or fairly applied.  

Mortgage application complexity

While the recent National Association of Realtors’ report Snapshot of Race and Home Buying in America did show some progress, the journey and barriers mean that there is still a chasm in the percentage of home ownership by various groups, including between different minority groups. The record high home prices, interest rates, and higher down payment requirements are problems, but these are amplified by credit scoring and a process that still relies on paper, faxes, and email. The process often ends in frustration and despair.   

For many, the mortgage loan process feels like a relic from a bygone era. Manual documentation and linear communication requirements are a testament to its inefficiency. Whether due to prior bankruptcies, immigration status, extenuating credit events, or multiple job changes, these applications are often set aside in favor of more “vanilla” profiles with a higher likelihood of closing. 

The move to AI

However, there is hope. AI and empathetic technology can help overcome this. Legacy processes that took weeks can now be completed in minutes to hours to help people make informed decisions. 

We’ve been able to eliminate human bias and increase mortgage application acceptance rates by up to 50% for some groups using augmented AI technology. AngelAi, an AI platform we created, is opening up the restrictive financial lending process. We reimagined the application and approval process, and AngelAi simplifies complex applications and delivers objective and accurate results, redefining trust. 

Augmented intelligence in financial services refers to the integration of artificial intelligence (AI) technologies with human expertise to enhance decision-making processes within the industry. Unlike artificial intelligence, which aims to replace human decision making entirely, augmented intelligence seeks to complement human intelligence, empowering financial professionals to make better-informed decisions faster and more accurately. 

Augmentation does not directly fight bias. The bias, intentional or not, comes from the process. The lending industry’s rigid workflow includes a series of documentation requirements and long, complex application procedures in English. This is intimidating to some in minority groups, discouraging them from participating in this mainstream financial undertaking. Our AI technology incorporates more than 100 languages and simplifies the loan application to a handful of easy questions, making the application process welcoming and inclusive.

Augmentation results are more exact and AI improves the applicant’s experience and ease of use, and reducing the applicant’s application costs. We put safeguards in place to ensure the safety, accuracy, and reliability of financial decisions, giving people a new level of confidence. The results are dramatic with approval rates improving for certain ethnic groups by over 50% in some cases.

Increase the home-buying pool

If we make access to financial resources easier and more equitable, we incentivize the developers. Currently, many developers prefer to build high-end condos when banks offer them below-market interest rates. With several recent large-scale banking collapses and other national economic problems, many of these condos or apartments sit empty, yet many still need housing. We’ve also allowed large investors to manipulate supply to artificially keep prices high with affordability out of reach. 

In the meantime, banks are not pursuing low income borrowers even though there are government programs designed specifically for them. While banks used to do a lot of these government loans, due to lending rule complexities, trained labor shortages, originated loans riddled with human error, and other problems, banks have avoided those products, restricting access to low-income financing.

Many myths and misperceptions keep eligible people out of the home-buying pool; even industry professionals lack all the knowledge about the government programs and policies designed to help. This is another area AI can address. Unbiased and immediate decisions and credit analyses can determine the lowest cost option for homeowners without relying on manual review, and it can be delivered over a simple chat discussion with an AI assistant.  

Reduce the intimidation and complexity

Minorities and people with disabilities are often intimidated by the process of speaking with a banking professional and therefore sequester themselves away from mainstream finance options. The non-judgmental interaction with an empathetic AI chatbot that speaks hundreds of languages is a necessary solution to dissolve these barriers. Fair lending is not just a platitude—it can be accomplished. 

We are committed to promoting access to fair lending to the groups that will experience the most benefit. AI tools can reduce the complexity of application processing and are agnostic to the specific loan product, so it doesn’t steer applicants to more expensive but more convenient loan products.

AI is also agnostic to the borrower’s complex financial circumstances. For example, it’s very common in minority communities to have multiple concurrent low-paying jobs and frequent job changes. Although the government guidelines clearly state that such income profiles are eligible, loan originators and underwriters are loathe to take those applications because of the immense amount of work required to review all of those documents. There’s also a high chance of making mistakes in those calculations and reviews. This leads to the borrowers being systemically neglected by the industry.

When people can have an improved experience with a loan application, gaining access to fair lending, they are empowered to better manage their financial future.

The importance of financial education

The other critical components are financial education and financial literacy, arguably the most important area for long-term progress in closing the wealth gap. This must start with basic education in schools and be developed and funded by financial institutions. While progress so far has been disappointing, we’re trying to solve it by having our AI program handle more direct questions and providing easy access to answers. 

We are engaging partners and stakeholders to work together to make the dream of home ownership more attainable thought AI-powered access, assistance, and education. But we need more focus and dialogue—what do you think? 

Pavan Agarwal is CEO of Celligence.

https://www.fastcompany.com/91129671/how-ai-creates-fair-lending?partner=rss&utm_source=rss&utm_medium=feed&utm_campaign=rss+fastcompany&utm_content=rss