June 26 – Introducing a Relational Framework for Less Discriminatory Alternative Search

Description: In the wake of the Administration’s recent executive order dismantling federal enforcement of disparate impact liability—a cornerstone of civil rights protections under the Fair Housing Act and Equal Credit Opportunity Act—there is an urgent need to explore alternative pathways to algorithmic fairness in housing and lending. This rollback threatens to erode decades of progress in combating systemic discrimination, particularly affecting communities of color, women, and other marginalized groups. As federal agencies deprioritize enforcement of disparate impact claims, the responsibility to ensure equitable outcomes increasingly falls on researchers, advocates, and industry leaders.

Join us for a timely webinar, “Introducing a Relational Framework for Less Discriminatory Alternative Search,” where we will unveil a novel approach to algorithmic fairness that emphasizes model selection across different families—termed ‘horizontal’ search—prior to within-model optimization. This framework offers a practical, resource-efficient method to identify fairer models, even in contexts with limited technical capacity. Through a case study in mortgage lending, we will demonstrate how this approach can lead to more equitable financial inclusion, providing actionable insights for practitioners, policymakers, and advocates committed to upholding fairness in an era of shifting legal landscapes.

Speakers:

Jameel Khan, Housing Market Data Scientist, NFHA

Dr. Michael Akinwumi, Chief AI Officer & Rita Allen Fellow, NFHA

Hana Samad, AI Governance Researcher, NFHA

Stephen Hayes, Partner, Relman Colfax PLLC

Brad Blower, Founder and Principal, Inclusive-Partners LLC

Miranda Bogen, Director, AI Governance Lab, Center For Democracy & Technology

Sasha Samberg-Champion, Special Counsel For Civil Rights, NFHA

Emmanuel Ogundimu, Associate Professor, University Of Durham, United Kingdom