On January 17, 2025, the CFPB issued a special edition of its Supervisory Highlights concerning advanced technologies. In particular, the Winter 2025 Supervisory Highlights Edition looks at select examinations of institutions that use credit scoring models, including models built with advanced technology commonly marketed as AI/ML technology, when making credit decisions.
Highlights of the report include the following findings by the examiners of the CFPB, particularly on credit scoring models and compliance with ECOA:
Credit card lenders’ use of complex credit scoring models. Examiners recently reviewed credit card lenders' use of AI or ML credit scoring models and found disparities in underwriting and pricing outcomes for Black, African American, and Hispanic applicants, along with inadequate compliance management systems. They instructed institutions to enhance testing protocols and use compliance tools suitable for their size, complexity, and risk to identify less discriminatory credit models.
Auto lenders’ credit scoring models. In examining auto lenders regarding credit scoring models, including those using AI/ML technology, Supervision found models utilizing over a thousand input variables, some being “alternative data” not traditionally linked to credit scoring. An Interagency Statement on Alternative Data in Credit Underwriting emphasizes that such unrelated data may pose consumer protection risks and require stronger compliance management.
Auto lenders’ adverse action notices. Recent examinations of auto loan originators revealed non-compliance with ECOA and Regulation B regarding adverse action notices. Examiners noted that institutions using AI/ML credit scoring models often failed to ensure proper selection of reasons for adverse actions based on these scores. Additionally, they did not validate their processes for selecting these reasons. To address these issues, examiners instructed institutions to test and validate their methodologies for identifying principal reasons in adverse action notices.
The Full Supervisory Highlights can be found here.