Air Date

April 20, 2022

Featured Guests

Chris Cartwright
President and CEO, TransUnion

Raghu Kulkarni
VP of Data Science, Discover Financial Services

Sammy Assefa
Head of AI and Machine Learning, U.S. Bank

Brian Stucky
Team Lead of Rocket Ethical AI, Rock Central

Moderators

David Hirschmann
Executive Vice President, New Strategic Initiatives, U.S. Chamber of Commerce, President and CEO, Center for Capital Markets Competitiveness (CCMC), President and CEO, Global Innovation Policy Center (GIPC), President and CEO, Chamber Technology Engagement Center (C_TEC)

Bill Hulse
Senior Vice President, Center for Capital Markets Competitiveness, U.S. Chamber of Commerce

Share

Financial institutions have always used consumer data like credit scores to inform their lending decisions. However, experts say that access to alternative data, combined with artificial intelligence, can help expand financial inclusion.

As financial institutions gain more in-depth information about consumers and their spending habits, they may be more willing to lend money to those who don’t have the same opportunities in the current financial system. To explore this relationship between data and AI in the financial sphere, the U.S. Chamber of Commerce Center for Capital Markets Competitiveness spoke with financial experts and leaders about how better data solutions can increase access to consumers and improve the economy.

Increased Access to Data May Create More Opportunities for Consumers

Financial institutions have been slow to embrace modern consumers’ spending habits. However, this began to change eight years ago when borrowers went from calculating a credit score based on an exact point in time to calculating it based on a three-year trajectory.

“This movement to trend data allowed [borrowers] to access more information about consumers and therefore score tens of millions more consumers here in the U.S.,” said Chris Cartwright, president and CEO of TransUnion.

Cartwright stated that if borrowers have access to alternative information on consumers’ payment habits, such as rent payments, utility bills, and wireless bills, they could provide even more inclusion opportunities.

“If provided on a consumer permission basis, all of that information has an important signal that can help the reporting agencies provide an accurate objective and even a more comprehensive picture of millions and millions more American adults,” said Cartwright.

Machine Learning Helps Institutions Make More Efficient Decisions

With the surplus of data these institutions receive each day, there needs to be a more efficient way to process it all. More inclusive financial services cannot happen if institutions are not equipped to handle increased applications and subsequent decisions. One way to efficiently work with this additional data is to utilize artificial intelligence and machine learning. AI can teach financial software applications to become more accurate at predictive financial tasks.

Brian Stucky, team lead of Rocket Ethical AI at Rock Central, talked about how his company has been embracing machine learning in multiple processes.

“Applying machine learning in all areas of marketing is something we have focused on very heavily,” said Stucky. “[It helps] us really tailor a client experience and … reach potential clients at the right time of their life, know what we think they're going to need, and then [determine] how we can best provide that to them.”

Banks Are Working With Regulators and the Government to Ensure Ethics

With financial institutions having increased access to various forms of consumer data, many worry about how these organizations plan to use it. Additionally, while AI and machine learning are meant to make processes more efficient and inclusive, some fear that the software will inhibit inclusivity.

Sammy Assefa, head of AI and machine learning at U.S. Bank, said how his team works with the government to ensure ethical practices regarding consumer data.

“It involves a review from a number of teams across the enterprise, including first [and] second line of risk, model governance, legal compliance, felony teams, you name it,” said Assefa. “That's in addition to the already existing rigorous review by the AI machine learning teams, so this will ensure that the models are explainable and unbiased.”

“We have a really strong partnership with the regulators, the OCC, [and have] AI governance meetings that run very regularly across the enterprise to make sure … all the stakeholders are aligned on the requirements, and [that] we adhere to them,” Assefa explained.