The CFPB released its latest fair lending report, Protecting consumers and encouraging innovation: 2019 Fair Lending Report to Congress with a covering blog posting.

Of note is a section, Innovations in access to credit with a subsection on “providing adverse action notices when using artificial intelligence and machine learning models.”  In this section, the Bureau wrote that “artificial intelligence (AI), and more specifically, machine learning (ML), a subset of AI…”, will be an area of where the Bureau will be monitoring for fair lending and credit access. As noted in a BallardSpahr blog,

The Bureau observes that there may be questions about how institutions can comply with ECOA and FCRA adverse action notice requirements ‘if the reasons driving an AI decision are based on complex interrelationships.’  It comments that ‘the existing regulatory framework has built-in flexibility that can be compatible with AI algorithms.’

There is regulatory flexibility to handle questions of AI and machine learning, according to the Bureau.  This was the message CDIA provided in a comment to the OMB earlier this year in connection with a January 2020 Request for Comments on a Draft Memorandum to the Heads of Executive Departments and Agencies, Guidance for Regulation of Artificial Intelligence Applications.

In the fair lending report, the CFPB said that the official commentary to ECOA’s Reg B says that when a creditor provides specific reasons for adverse actions, that commentary provides that in giving specific reasons for adverse action,

a creditor need not describe how or why a disclosed factor adversely affected an application, or, for credit scoring systems, how the factor relates to creditworthiness.  Thus, the Official Interpretation provides an example that a creditor may disclose a reason for a denial, even if the relationship of that disclosed factor to predicting creditworthiness may be unclear to the applicant.”  The Bureau comments that this flexibility could be useful to creditors when issuing adverse action notices “based on AI models where the variables and key reasons are known, but which may rely upon non-intuitive relationships.

The Bureau adds that “neither ECOA nor Regulation B mandate the use of any particular list of reasons.  Indeed, the regulation provides that creditors must accurately describe the factors actually considered and scored by the creditor, even if those reasons are not reflected on the current sample forms.”  This  flexibility could be useful to creditors “when providing reasons that reflect alternative data sources and more complex models.”

Artificial intelligence, according to the Bureau, “hold[s] great promise…to facilitate use of AI for credit underwriting compatible with adverse action notice requirements.”  The BallardSpahr blog points out that despite the flexibility of the existing regulatory framework “there still may be some regulatory uncertainty about how aspects of the adverse action notice requirements apply in the context of AI/ML.”  The Bureau encourages entities to consider using the Bureau’s new innovation policies (e.g. No-Action Letter Policy) to address potential compliance issues.