Credit Inequality and Artificial Intelligence
By: Geff Woodward
In the land of opportunity, where dreams are said to come true, there exists a persistent and troubling issue that affects millions of Americans: credit inequality. While the United States prides itself on its economic prosperity and upward mobility, the harsh reality is that access to credit is far from equitable. In this blog, we'll delve into the intricacies of credit inequality in America, exploring its root causes, consequences, and potential solutions.
Credit inequality, at its core, refers to disparities in access to affordable credit and financial services among different groups of people. It's important to note that credit inequality is closely intertwined with other forms of economic inequality, such as income and wealth inequality. One of the primary factors contributing to credit inequality is the reliance on credit scores to determine creditworthiness.
Those with limited or poor credit histories often find themselves excluded from favorable lending terms, making it difficult to access credit for important life events like buying a home or starting a business. Systemic Discrimination also plays a part in this inequality disparity. Discrimination based on race, gender, and ethnicity is a deep-seated issue that affects credit access. Studies have shown that historically marginalized groups face greater obstacles in obtaining credit, even when controlling for other factors.
This discrimination may present when seeking loans or mortgages. A recent study of nearly 7 million 30-year mortgages by the University of California at Berkeley found that African-American and Latino applicants were charged higher interest when compared with White borrowers. But the most concerning thing is that, even when they applied online, minorities still ended up paying higher rates. Somehow, lending institutions have been able to reproduce that discrimination in software-based lending.
The issue seems to lay on the fact that despite Artificial Intelligence (AI) and Machine Learning (ML), online home lending applications were built from old mortgages that were already biased. Programmers loaded in large data sets to teach the system how to respond to new information, and then it identifies historical patterns within the data that will be used to make future lending decisions.
Given all the above, it would not surprise anyone that another study that analyzed 10 million mortgage applications from Home Mortgage Disclosure Act (HMDA) data, showed thatAfrican-Americans had the highest denial rates for mortgages in 2018 at 17.4% while White borrowers had the lowest at 7.9%. We are depriving people of one of the basic human rights: a home.
More than ever, given the current financial and real estate situation, there is a pressing need for an unbiased financial score that can objectively and holistically measure consumers' financial heath from scratch.
AI is not inherently perpetrating credit inequality, but it can inadvertently exacerbate existing disparities if not used and regulated responsibly. The impact of AI on credit inequality largely depends on how it is implemented and the data it relies on.
AI itself is a tool, and its impact on credit inequality depends on how it is designed, implemented, and regulated. While AI has the potential to make lending more efficient and inclusive, it also carries the risk of perpetuating existing inequalities if not used responsibly and with careful consideration of its potential biases. Balancing technological innovation with ethical and equitable practices is essential to ensure that AI benefits all members of society.