Senators Urge Regulatory Oversight of AI to Prevent Discrimination

As auto lenders increasingly turn to technology like artificial intelligence to improve business practices, two senators are pressing regulators to ensure anti-discrimination laws keep up with the innovation.

Sen. Elizabeth Warren (D-Mass.) and Doug Jones (D-Ala.) sent a letter to the leaders of the Consumer Financial Protection Bureau, Federal Reserve, Office of the Comptroller of the Currency, and the Federal Deposit Insurance Corp. requesting answers to what the agencies are doing “to identify and combat lending discrimination by lenders who use algorithms for underwriting,” the letter notes.

To support their request, Warren and Jones presented research in the letter that examined whether differences between human and algorithmic decision-making would influence whether to issue loans and at what terms. The research found that “algorithmic lending” reduced the likelihood that the loan applications for borrowers of color are rejected, reducing discrimination in a lender’s decision-making process.

Conversely, the research results also found lenders leveraging algorithms in underwriting charge minority borrowers “6 to 9 basis points higher interest rates” — which is almost identical to the interest rate premium that minority borrowers face with traditional underwriting — coming to the conclusion that the algorithms used by fintech lenders are “as discriminatory as loan officers,” the senators explained.

Auto financiers leveraging artificial intelligence in their operations include Tricolor Auto Acceptance, which integrated an AI algorithm into its underwriting model earlier this year. Though, the algorithm, which was tested for eight months before it launched, effectively “sits on top” of the current risk model, creating an additional layer of “cleaner segmentation,” Chief Executive Daniel Chu previously told AFN.

While the Senators are aware of the “opportunity of algorithmic underwriting’s potential to reduce discrimination,” the letter notes, one lender told AFN there is a shortcoming with the technology.

“Artificial intelligence can be discriminatory in terms of underwriting because it is difficult to prove how you came up with an answer — you can’t quantify what you are doing,” said Micky Watts, SVP at Anderson Brothers Bank.

The Senators set a June 24 deadline for answers to a series of five questions. Click here to view the full letter issued June 10.