CFPB Reminds Lenders that Algorithms Are Not a Shield for Discrimination

  • The CFPB released Consumer Financial Protection Circular 2022-03 and an accompanying press release detailing the agency’s position on the application of anti-discrimination laws to consumer lenders relying on so-called “black box” credit models that use complex algorithms or artificial intelligence to evaluate creditworthiness.
  • Specifically, the Circular addressed the application of the Equal Credit Opportunity Act (ECOA) and Regulation B to credit decisions made through black box models. The ECOA and Regulation B require that creditors provide consumers with a “specific” statement that “indicate[s] the principle reason(s) for the adverse action.” Adverse actions can include, among other things, denying an application for credit, terminating an existing credit account, changing the terms of an existing account, or refusing to increase a credit limit.
  • In the Circular, the CFPB asserted that a creditor cannot justify noncompliance with the ECOA/Regulation B adverse action notice requirements based on the fact that the technology it uses to make such decisions is too complex or opaque to understand. Companies can still be liable for failing to provide specific and accurate reasons for an adverse action, and, according to the CFPB, a creditor’s lack of understanding of its black box credit model is not an adequate defense against liability for noncompliance with the ECOA and Regulation B.