Algorithmic Discrimination Beyond Employment: Credit, Housing, Tenant Screening, and Insurance Models Under Legal Fire

Eric P. Sublett, Partner | Relman Colfax PLLC

Eric P. Sublett is a Partner at Relman Colfax PLLC, where he has practiced in the firm’s civil rights counseling group since 2012. He primarily counsels financial institutions on best practices for fair lending compliance, and much of his work addresses algorithmic fairness and the discrimination risk posed by statistical and machine learning models. He helps clients identify and reduce those risks in traditional lending as well as in the financial technology and online platform settings where they increasingly arise. He is especially well suited to this program: he has worked on the firm’s independent fair lending monitorship of Upstart Network’s lending model and on independent civil rights audits, bringing a hands-on view of how a model is tested, documented, and held accountable.

• Education & Credentials

Mr. Sublett received his J.D. cum laude from Harvard Law School. He also holds an M.A. from Yale University and a B.A. with highest honors from The University of Texas at Austin. He is admitted to practice in the District of Columbia and Connecticut.

• Recognition & Leadership

Mr. Sublett’s featured matters at Relman Colfax include the firm’s service as independent fair lending Monitor of Upstart Network’s lending platform, an appointment made under a December 2020 agreement among Upstart, the NAACP Legal Defense Fund, and the Student Borrower Protection Center, which produced a series of periodic reports evaluating the platform’s fair lending implications. His featured work also includes the independent civil rights audit of State Street Corporation, whose final report was published on May 30, 2023, and the firm’s Civil Rights Audit of Facebook.

• Professional Involvement

Mr. Sublett has spoken at conferences organized by the Business Law Section of the American Bar Association and by the National Community Reinvestment Coalition, and he has appeared as a guest speaker at Howard University School of Law. He previously served as a visiting fellow with the Information Society Project at Yale Law School.

• Experience

Before joining Relman Colfax, Mr. Sublett handled fair housing matters as a volunteer attorney with the Connecticut Fair Housing Center. At the firm, his counseling practice centers on fair lending, with particular emphasis on how statistical and machine learning models can create discrimination risk and how that risk can be detected and mitigated.

Sacha B. Markano-Stark, Counsel | Relman Colfax PLLC

Sacha B. Markano-Stark is Counsel at Relman Colfax PLLC, which she joined in 2021, and she practices in both the firm’s civil rights counseling and litigation groups. She advises financial institutions and other organizations on compliance and best practices in fair housing, fair lending, and fair banking. Her counseling also covers tech bias, appraisal-related issues, financial equity, and affirmative lending programs, while her litigation work challenges redlining, appraisal discrimination, and other discriminatory practices. She is especially well suited to this program: she handles Fair Housing Act and ECOA matters on both the compliance and litigation side and has presented on algorithmic bias in housing-related transactions, bringing both a counseling perspective and a courtroom view of how discrimination claims against a model are built.

• Education & Credentials

Ms. Markano-Stark earned her J.D. from New York University School of Law, and she also holds an M.A. from New York University and a B.A. from Reed College. She is admitted in the District of Columbia and Virginia (inactive).

• Recognition & Leadership

Ms. Markano-Stark served as a judicial clerk at both the federal trial and appellate levels, clerking for the Hon. Martha Craig Daughtrey of the U.S. Court of Appeals for the Sixth Circuit and the Hon. Paula Xinis of the U.S. District Court for the District of Maryland.

• Professional Involvement

Ms. Markano-Stark is a co-author of “Adding Robust Consideration Of Race To Community Reinvestment Act Regulations: An Essential And Constitutional Proposal,” published by the National Community Reinvestment Coalition in 2021. She was also one of the presenters of “Mining the Data: Algorithmic Bias in Housing Related Transactions,” an NFHTA Forum Online program held on January 19, 2022.

• Experience

Before joining Relman Colfax, Ms. Markano-Stark was a litigation associate at Paul, Weiss, Rifkind, Wharton & Garrison LLP, where she handled complex civil litigation, internal investigations, corporate compliance, anti-corruption and anti-bribery, and data security matters while maintaining an active pro bono practice. During law school she worked with the Equal Justice Initiative in Alabama, the Promise of Justice Initiative in Louisiana, and the U.S. Department of Justice’s Civil Rights Division, Special Litigation Section, and Criminal Division, Public Integrity Section. She also brings significant experience in voter protection and voting rights.

SESSION 1 – Where the Liability Comes From: Disparate Impact and the Consequential-Decision Model in Credit, Housing, and Insurance | Sacha B. Markano-Stark | 12:00pm – 1:00pm

This session maps the liability landscape for algorithmic decision systems outside employment. Attorneys will learn how the disparate impact framework applies to models under the Fair Housing Act, ECOA and Regulation B, and the state statutes reaching insurance and consumer scoring, including the burden-shifting structure, the business-justification defense, and the less-discriminatory-alternative battleground. The session examines what Louis v. SafeRent Solutions established about vendor exposure for tenant-screening models, the adverse action notice obligations that apply when a model makes or informs the decision, and the state insurance rules requiring insurers to test algorithms and external data for unfair discrimination. Attendees will leave with a practical approach to allocating model risk by contract, handling the discovery fight over source code and training data, and counseling clients who license third-party scores they cannot see inside.

BREAK | 1:00pm – 1:10pm

SESSION 2 – Testing, Challenging, and Defending the Model: Fairness Metrics, Less Discriminatory Alternatives, and the Audit Record | Eric P. Sublett | 1:10pm – 2:10pm

This session goes inside the model to show how algorithmic fairness is measured, attacked, and defended. Attorneys will learn what adverse impact ratios, calibration, and error-rate disparities actually measure and where they conflict, how disparate impact testing is run in practice, including proxy analysis when protected-class data is unavailable, and how less discriminatory alternative searches are used by plaintiffs, regulators, and defendants. The session closes with the governance side: model documentation, independent testing, audits and monitorships, and remediation when a test finds a problem. Attendees will leave knowing how to structure model testing under privilege, translate quantitative evidence for judges and juries, and assemble a model file that holds up when the regulator or the plaintiff’s expert comes calling.