August 12, 2019
Partner Jean Noonan comments for Bloomberg on HUD proposal to redefine use of disparate impact
On August 9, Bloomberg Law reported on how the Department of Housing and Urban Development's (HUD) proposed Disparate Impact Rule under the Fair Housing Act could affect lenders' use of machine learning, artificial intelligence, and other technology.
According to Bloomberg, the HUD proposal would allow bias claims to be dismissed if lenders can prove any of three defenses. The first would be that the inputs to an algorithmic model aren't substitutes for prohibited characteristics such as gender or race, and that the model is actually predictive of risk. The second would be to show that a "recognized third party" created the algorithm; and the third would be to show that a neutral third party had analyzed an algorithmic model and determined it was statistically sound, predictive of risk and didn't use inputs that substitute for protected characteristics.
"The good thing about this rule is that it recognizes that algorithms and credit models are typically well designed to predict risk and creates a complete defense, if you meet the standard," said Hudson Cook Partner Jean Noonan. However, Jean noted that some of HUD's terms are vague, particularly around what factors may be "substitutes or close proxies" for protected classes. Many algorithmic models consider factors that may be correlated to protected classes like gender or ethnicity, Jean said. But the correlation with a protected class can't in itself be a violation of the Fair Housing Act. HUD would need to define the degree to which a correlation of a factor, such as geography, for a protected class, such as race, would be deemed significant or not.
Jean is a partner in the firm's Washington, DC office. She advises clients on consumer financial services, fair lending, marketing, financial privacy and consumer protection matters. She counsels financial institutions and others in complying with laws related to consumer credit, privacy, telemarketing and unfair trade practices. Jean also represents clients in government investigations, examinations, and enforcement actions before federal agencies, including the Consumer Financial Protection Bureau, Federal Trade Commission, and federal prudential regulators and in other ancillary matters.
Subscribers to Bloomberg may click here to read the article.