To better address inequities in the backend technologies in the mortgage lending business, a leading association is hosting an event where participants will brainstorm solutions.

The Tech Equity Hackathon, scheduled to take place on June 16-18 in Washington, DC, is the brainchild of the National Fair Housing Alliance, an association of civil rights and fair-housing organizations. According to NFHA Chief Tech Equity Officer Dr. Michael Akinwumi, algorithms dealing with mortgage underwriting and appraisals can often be biased against people of color buying homes, and undervalue homes in neighborhoods with high proportions of people of color. For instance, a 2021 investigation conducted by reporting outlet the Markup found that due to AI bias, Black mortgage applicants were 80 percent more likely to be rejected than white applicants; in Waco, Texas, Latino applicants were 200 percent more likely to have their applications rejected by lenders. 

It’s really important that the people developing the solutions also represent the consumers or demographics that are affected.

NFHA Chief Tech Equity Officer Michael Akinwumi

Improving diversity and fairness in homeownership, Akinwumi said, requires diverse groups working on the problem.

“It’s really important that the people developing the solutions also represent the consumers or demographics that are affected,” he said. “Some financial institutions, when you look into the teams that are developing the models, their racial profiles don’t really represent the people they’re trying to serve.”

The hackathon concept has been in development at NFHA since its Tech Equity Team was created in 2020, but was stalled during the pandemic before being picked up again for development last year. Applicants to the hackathon are asked to choose to focus on one of three “problem statements” developed by NFHA: how to develop a valuation model of homes “regardless of the racial composition of the area”; how to develop a model for identifying disparities in mortgage underwriting; and how to develop tools to mitigate bias in housing and lending algorithmic models.

To attract a diverse pool of applicants, Akinwumi said, NFHA has reached out to computer science departments at historically black colleges and universities (HBCUs) and partnered with UrbanGeekz, a news website focused on underrepresented communities in the tech industry. As a public event, the hackathon will help raise awareness about some of the housing and lending issues it’s designed to address. But Akinwumi says it can also help NFHA develop its policy strategies, and attract more diverse pool of professionals focused on AI and machine learning in the housing sector.

“We’re hoping that at the end of the hackathon, we’ll have additional ideas from stakeholders and policymakers, especially in education policy to make sure that those who are pursuing careers in data science or AI are diverse,” he said. “That is the problem we’re trying to solve.”

This year’s hackathon has been promoted mainly to college students and early-career tech workers. Moving forward, Akinwumi said NFHA will look to attract an even younger cohort of participants. 

“We’re planning to take this to high schools,” he said. “Based on the application data we have, we think there’s an opportunity to campaign there. That’s when people decide what they want to major in in college, and we want to encourage people to specialize in AI and take prerequisites based on it.”

(Roman Didkivskyi/iStock)