On November 3, 2025, Skadden and Troutman co-hosted a conference in Washington, D.C. titled “Fair Lending 2025: Navigating Turbulent Waters.” Leading the conference were Anand Raman, head of Skadden’s Consumer Financial Services practice, and Chris Willis, co-leader of Troutman’s Consumer Financial Services Regulatory Practice Group. Panelists also included Lori Sommerfield, partner at Troutman, Darren Welch, partner at Skadden, Bryan Burcat, associate at Skadden, and Lane Page, associate at Troutman.
The conference included panels on (1) The New Landscape of Federal and State Fair Lending Supervision and Enforcement, (2) Disparate Impact: Past, Present and Future, (3) Hot Topics in Fair Lending and (4) Fair Lending Issues in Model Development and AI.
Opening Remarks
Mr. Raman and Mr. Willis commenced the conference by describing the current state of consumer finance in light of Trump administration priorities in areas such as DEI, disparate impact, AI, debanking, and deregulation, as well as significant changes in agency size and activity level. They commented that the landscape of fair lending has shifted dramatically, with 2025 being one of the most eventful years ever for fair lending, resulting in significant regulatory uncertainty.
Panel 1: New Landscape of Fair Lending Oversight
The panel discussed implications of the Trump administration’s approach to a number of issues, the states’ responses, and practical steps for mitigating risk.
- There have been significant changes at the federal banking and regulatory agencies. Overall, federal agencies have rolled back fair lending enforcement and pivoted to focusing on new topics, including debanking and special purpose credit programs.
- State regulators and attorneys general have significant authority in the fair lending space. States have historically focused on indirect auto financing and redlining (which is assessed at the local community level) and in some instances compete with one another to be the first to bring enforcement actions. In light of the federal pullback from fair lending enforcement, fair lending scrutiny at the state level is building. The staffing reductions at federal agencies have also resulted in former federal enforcement attorneys joining state enforcement offices.
- Actions today by the federal government may impede or delay the enforcement abilities of a future administration, including expected changes to Regulation B as well as dramatic staffing cuts across agencies. However, absent statutory changes by Congress, these changes may be temporary. (Note: On November 13, 2025, the CFPB issued a proposed rule for comment that would significantly amend Regulation B.)
- Companies would be well advised to stay the course and maintain or build strong fair lending compliance management systems, particularly in light of risk from private and state action as well as future administrations given long statutes of limitations.
Panel 2: Disparate Impact: Past, Present and Future
The panel explored federal executive and agency actions aimed at eliminating disparate impact liability, recent disparate impact case law and enforcement activity, and practical aspects of assessing and mitigating disparate impact risk.
- Executive Order 14281 (Restoring Equality of Opportunity and Meritocracy) broadly calls for federal agencies to eliminate the use of disparate-impact liability. Regulatory agencies have updated their examination manuals and guidance to omit references to disparate impact.
- The U.S. Supreme Court will soon be deciding whether to hear the Emigrant Bank case, where the Second Circuit affirmed the use of jury instructions permitting plaintiffs to prove disparate impact by establishing that one racial group was adversely affected, rather than requiring a showing that the impact was disproportionate compared to other groups. Also at issue are use of a “fairness-based” test for equitable tolling and establishing a “robust causality” pursuant to the Inclusive Communities case based on statistics alone.
- HUD’s disparate impact regulation under the Fair Housing Act (FHA) has evolved over time as administrations have shifted between Republican and Democratic control. Recently, HUD sent a final rule on its implementation of the FHA disparate impact standard to the Office of Management and Budget, which may face Administrative Procedure Act challenges.
- In the Trump administration, reverse discrimination is a real concern. In conducting disparate impact testing, companies should be mindful of disparities disfavoring groups that have historically been viewed as advantaged.
- Some regulators and courts have inferred intent from statistical disparities, thus blurring the lines between disparate impact and disparate treatment.
- Redlining is a classic example of a theory that legally has been viewed as disparate treatment but which in practice regulators treat as disparate impact, with little or no evidence of intent in many enforcement actions.
- Disparate impact remains the law under the FHA and Equal Credit Opportunity Act (ECOA) as interpreted by courts, as well as under many state laws. States have also issued guidance about disparate impact under existing law and introduced legislation to codify disparate impact.
- Given present and future legal risks, lenders should consider continuing disparate impact fair lending testing.
Panel 3: Hot Topics in Fair Lending
This panel focused on three leading hot topics: (1) debanking, (2) redlining and (3) special purpose credit programs (SPCPs).
- Debanking initiative: Pursuant to Executive Order 14331 (Guaranteeing Fair Banking for All Americans), current federal policy prohibits denying access to financial services based on constitutionally or statutorily protected beliefs, affiliations or political views. In a sharp rebuke to “Operation Chokepoint” efforts in prior administrations, Executive Order 14331 mandates that regulators review their supervisory and complaint data to identify institutions that have engaged in unlawful debanking based on religion. Some states have also passed debanking laws.
- Importantly, section 5(c) of Executive Order 14331 requires that federal banking regulators review their supervisory and complaint data to identify any institution that has engaged in unlawful debanking on the basis of religion in violation of ECOA. If an institution is unable to achieve compliance, then the agency must refer the matter to the attorney general for appropriate civil action. This provision poses significant risk to financial institutions.
- We have already seen swift action by federal agencies to implement Executive Order 14331 and curtail debanking practices. The federal banking agencies have removed reputational risk from their guidance and issued extensive information requests to institutions related to debanking. The Small Business Administration has likewise issued information request letters to thousands of lenders. Debanking concerns must be balanced with Bank Secrecy Act/anti-money laundering and safety and soundness risks.
- Redlining: The Trump administration has indicated that it does not intend to pursue redlining enforcement at the federal level, and a number of consent orders have been terminated. However, we expect states to increase redlining enforcement.
- The Section 1071 Small Business Lending Rule will create risks of redlining litigation from private plaintiffs, states, and/or a future administration once data becomes available. Institutions may wish to use this period while Section 1071 requirements are paused and being revised to build out their small business fair lending risk management practices. (Note: On November 13, 2025, the CFPB issued a proposed rule for comment that would significantly revamp the Section 1071 rule.)
- SPCPs: SPCP risk is rising, despite ECOA and current regulations permitting these programs. Person-based programs in particular have significant constitutional and other risks. Geographical SPCPs in the mortgage context remain common, but some constitutional and statutory risk remains. The panel discussed a number of alternatives to SPCPs, including lending programs that consider other factors as well as affirmative advertising. (Note: The proposed rule to amend Regulation B, issued by the CFPB on November 13, 2025, would place significant limits on the use of SPCPs by for-profit organizations.)
Panel 4: Fair Lending Issues in Model Development and AI
The final panel of the conference discussed the uses and risks of AI, federal and state regulation, recent enforcement activity, and adverse action issues.
- AI is being increasingly used in the financial services industry and has significant potential to improve customer access and experience, risk management and efficiency, but the use of AI presents fair lending risks, including with respect to adverse action notices.
- There is no comprehensive federal AI law or regulation; the regulation that does exist is a patchwork of state laws and guidance. Colorado has passed the most comprehensive state law pertaining to the use of AI in financial services. Some other states have issued guidance stating that AI models are subject to state UDAP laws.
- A recent Massachusetts enforcement action faulted a lender for failing to perform disparate impact testing of its AI models, use of discretionary human decision-making in connection with training data, use of certain variables including cohort default rate, and inadequate adverse action notices.
- CFPB guidance issued under former Director Chopra’s leadership indicated aggressive oversight and scrutiny of AI, but much of this guidance has since been withdrawn. As a result, the Bureau’s position on the use of AI is unclear at this point.
- With respect to generative AI and chatbots, fair lending risk mitigation may include using data sets that are large and representative with pre- and post-deployment monitoring. A key risk with AI chatbots is their potential for differential treatment of individuals based on how they speak, write or otherwise interact with the bot, such that the bot makes assumptions about the individuals and uses this information as the basis for their product offerings or assistance. Chatbots can be evaluated and monitored similar to customer service representatives.