On March 29, the Federal Reserve Board, the Consumer Financial Protection Bureau, the Federal Deposit Insurance Corporation, the Office of the Comptroller of the Currency, and the National Credit Union Administration (the "Federal Agencies") issued a request for information ("RFI") from financial institutions, trade associations, consumer groups, and other stakeholders on the financial industry's use of artificial intelligence ("AI"). The RFI broadly seeks insight into the industry's use of AI in the provision of financial services to customers and appropriate AI governance, risk management, and controls. While the RFI should not come as a surprise (for several years, regulators have highlighted the growing use of AI and machine learning by financial institutions and technology firms), it is the most coordinated effort to date by the Federal Agencies to better understand the potential benefits and risks of AI. It follows a speech earlier this year in which Federal Reserve Board Governor Lael Brainard previewed the potential for additional "supervisory clarity" in this area.
Risks and Rewards
The Federal Agencies acknowledge in the RFI the importance of AI to the industry and its customers, including with respect to AI's use in flagging unusual transactions, personalization of customer services, credit decision-making, risk management, textual analysis (handling unstructured data and obtaining insights from that data or improving efficiency of existing processes), and cybersecurity. The RFI also notes the potential safety and soundness risks of AI, including operational vulnerabilities, cyber threats, information technology lapses, third-party risk, and model risk. Consumer risks are also identified, such as risks of unlawful discrimination, unfair, deceptive, or abusive acts or practices, and privacy concerns. In addition, the RFI discusses the importance of "explainability," which refers to "how an AI approach uses inputs to produce outputs." Some AI approaches exhibit a "lack of explainability" for their overall functioning or how they arrive at individual outcomes, which can give rise to challenges in legal compliance, audit, and other contexts.
Request for Information
The RFI seeks comment on the following areas:
- risks from broader or more intensive data processing and usage;
- "overfitting," which occurs when an algorithm "learns" from idiosyncratic patterns in the training data that are not representative of the population as a whole;
- cybersecurity risk;
- "dynamic updating," which refers to AI's ability for it to learn or evolve over time as it captures new training data;
- AI use by community institutions;
- oversight of third parties that have developed or provide AI; and
- fair lending.
Fair lending appears poised to be a central supervisory concern of the Federal Agencies when evaluating AI design and usage. More questions are posed concerning fair lending than any other area in the RFI. In particular, the Federal Agencies seek input on the following questions:
- What techniques are available to facilitate or evaluate the compliance of AI-based credit determination approaches with fair lending laws or mitigate risks of non-compliance?
- What are the risks that AI can be biased and/or result in discrimination on prohibited bases? Are there effective ways to reduce risk of discrimination, whether during development, validation, revision, and/or use? What are some of the barriers to or limitations of those methods?
- To what extent do model risk management principles and practices aid or inhibit evaluations of AI-based credit determination approaches for compliance with fair lending laws?
- What challenges, if any, do financial institutions face when applying internal model risk management principles and practices to the development, validation, or use of fair lending risk assessment models based on AI?
- What approaches can be used to identify the reasons for taking adverse action on a credit application when AI is employed? Do existing rules under the Equal Credit Opportunity Act provide sufficient clarity for the statement of reasons for adverse action when AI is used?
The RFI reflects an increasing interest in AI by the Federal Agencies, especially as it relates to the risks posed to consumers and the safety and soundness of financial institutions. We remain attentive to trends and developments in this ever-evolving space and would be happy to discuss any questions or concerns with respect to the use of AI in financial services.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.