Skip to main content

Adverse Action Notice Requirements when using Artificial Intelligence

On September 19, 2023, the CFPB published a circular on the proper use of Regulation B (“Reg B”) sample forms and adverse action notification requirements. With the rise of artificial intelligence and machine learning, a question was posed to the CFPB if when using artificial intelligence, algorithms, or complex credit models, may the sample checklist of reasons for the adverse action be utilized even when the sample reasons do not relate to the specific reasons for denial. 

The Equal Credit Opportunity Act (“ECOA”) and Reg B require that an adverse action notice must be provided to any applicant who applies for credit and whose creditor denies the credit request or does not offer the credit terms requested by the applicant. The CFPB provided sample forms, included in Reg B, to make adverse action notices, and various adverse reason codes, uniform. The ECOA and Reg B require that the reasons be specific and must be the principal reasons for the action taken. This information is utilized to identify discrimination and promote fairness pursuant to the ECOA. These forms also assist consumers in taking steps to correct mistakes or improve their credit files by being clear and concise in the reasons the credit was denied or not approved as requested. 

In its response, the CFPB stated that because artificial intelligence may pull data that is not readily available in a borrower’s credit file and utilize data pulled from other sources, the predefined reasons on the sample adverse action forms provided under Reg B may not accurately reflect the specific reasons the borrower received an adverse decision.  This is in addition to the risk posed by use of such source material which may not properly represent the borrower’s ability or willingness to repay their credit obligations. As a creditor is required to provide specific reasons for credit denial, reliance on the predefined general reasons for denial may result in noncompliance with ECOA and Reg B requirements in cases where a creditor relied on an algorithm which includes artificial intelligence, such as through consumer surveillance. 

As a result of use of information outside of the consumer report obtained from a consumer reporting agency, the CFPB indicated a creditor may be out of compliance with ECOA and Reg B if the creditor relies only on the predefined reasons on the sample forms.  The CFPB stated the checklist of reasons provided in Reg B is simply an example and should not be utilized as a template to satisfy the obligations under EOCA for adverse action notices. The reasons a loan is being adverse actioned must properly represent the borrower and accurately represent the information contained in the loan file for that borrower. Lenders may also not be overly broad or vague with the reasons used to shroud the accurate or detailed adverse action reason. The sample forms do show common reasons adverse action is taken but those reasons may not be appropriate for all borrower or for all creditors.

DocMagic will continue to provide CFPB updates as they become available.  If you have additional questions about this topic, please contact DocMagic’s Compliance Department.

SOLUTIONS THAT WORK. TECHNOLOGY TO STAY COMPLIANT.