CFPB Says Lenders Must Explain AI-Driven Credit Denials

The Consumer Financial Protection Bureau (CFPB) has released guidance for lenders who utilize artificial intelligence (AI) and complex models in their credit decision-making processes.

The guidance emphasizes the need for lenders to provide specific and accurate explanations when making credit denials, the regulator said in a Tuesday (Sept. 19) press release.

With the growing use of advanced algorithms and personal consumer data in credit underwriting, the reasons for adverse credit actions have become more diverse, according to the release. Lenders often rely on complex algorithms and AI technologies to make predictive decisions in their underwriting models.

These algorithms are fed with large datasets, sometimes including data obtained from consumer surveillance, the release said. Consequently, consumers may be denied credit based on reasons that they may not consider relevant to their financial situation.

To address this issue, the CFPB emphasizes that creditors must be able to clearly explain their reasons for denial in a specific and accurate manner, per the release. Merely using sample adverse action forms and checklists provided by the CFPB is insufficient if they fail to reflect the actual reason for the denial of credit or a change in credit conditions. This requirement applies even when complex algorithms and black-box credit models are used, making identifying the specific reasons for adverse actions challenging.

The CFPB’s guidance clarifies that creditors cannot rely on broad reasons or select factors from a checklist to explain adverse actions, according to the press release. Instead, they must provide detailed explanations that go beyond general terms. For instance, if a creditor lowers a consumer’s credit line based on behavioral spending data, the explanation should include specific negative behaviors that led to the reduction rather than a general reason like “purchasing history.”

The CFPB stresses the importance of creditors ensuring compliance with the law by disclosing the specific reasons for adverse actions. This requirement holds true even if consumers may be surprised or upset to learn that their credit applications were evaluated based on data that may not intuitively relate to their finances, the release said.

The CFPB firmly states that the Equal Credit Opportunity Act mandates creditors to provide accurate and specific reasons for adverse actions, regardless of the complexity of the algorithms or models used, per the release.

This guidance comes about 18 months after the CFPB published a reminder for consumers that federal law requires companies to provide them with specific reasons for denying a credit application, even when using a “black-box model” to make decisions.

“The law gives every applicant the right to a specific explanation if their application for credit was denied, and that right is not diminished simply because a company uses a complex algorithm that it doesn’t understand,” CFPB Director Rohit Chopra said at the time.