Mapping the Landscape: Antidiscrimination Law and AI
This is a table-setting session that provides a high-level overview of the evolving relationship between AI and antidiscrimination law. The core tenants of antidiscrimination law currently focus on intent and formal classification, yet these tenants are likely to become more marginalized while disparate impact theory will move to the fore as AI pervades life and decision-making. The panel will map this terrain, identify key concepts at a broad level of generality, surface points of friction, and note emerging areas of alignment. The panel will touch on whether lawyers and technologists share similar conceptions of discrimination, measurement, and proof. Because both the technology and the legal landscape are rapidly changing, the session will consider how to think about equality in such a dynamic environment.
Foundations of AI and Antidiscrimination Law
These sets of sessions provide a technical and practical understanding of both the technology and the law. The focus will be contemporary AI systems, their design, and their limitations, as well as domain-specific legal frameworks and doctrines. The sessions will bridge legal and technical perspectives to equip non-technical participants with appropriately detailed knowledge of the technology and to equip non-legal technologists with a nuanced understanding of the workings and goals of antidiscrimination law. Part I consists of two instructional sessions that serve as the convening’s primary instructional component, where attendees will learn more directly about antidiscrimination law, AI technology, and how they intersect. Part II consists of facilitated discussion groups designed to be more interactive and practical.
Litigation: Possibilities and Limits of Existing Doctrines
This session will survey the emerging landscape of AI discrimination litigation, from the DOJ’s landmark settlement with Meta (the first federal case challenging algorithmic ad delivery under the Fair Housing Act), to the recent Mobley v. Workday collective action AI vendor liability in employment case. This session will consider the possibilities and limits of litigation with respect to AI and machine learning. It will explore how plaintiffs and courts are grappling with traditional antidiscrimination frameworks, with special attention to disparate impact. The aim is to consider the strengths and weaknesses of the application of current law to algorithmic decision-making.
Case Study Working Groups
This session uses hypothetical case studies—many based on real-world disputes—to explore how AI might reshape, confound, or challenge the enforcement of antidiscrimination law in a variety of sectors. Breakout participants will grapple with how AI systems may perpetuate or obscure discrimination based on a range of protected characteristics, including race, gender, disability, language, national origin, religion, and family status. The session pushes attendees to think strategically about both legal frameworks and technological design choices, examining how discrimination is identified, proven, and addressed in AI contexts. Through interdisciplinary discussion, participants will consider how antidiscrimination law and AI governance may need to evolve to meet these emerging challenges.
Corporate Governance and Industry Norms
As artificial intelligence and emerging technologies are increasingly embedded in decisions affecting employment, housing, credit, healthcare, surveillance, and access to opportunity, corporate norms and internal “best practices” shape the definition, application and enforcement of antidiscrimination protections. This panel examines how companies do and should operationalize concepts such as fairness, bias mitigation, transparency, and accountability—and how those choices influence litigation, regulatory oversight, and policy development. Panelists will explore the extent to which voluntary standards, audits, and governance frameworks might meaningfully advance (or undermine) antidiscrimination goals.
Creating New Law: Regulations and Legislation
This session examines emerging approaches and continuing challenges to regulating AI discrimination, exploring international, federal, state, and city initiatives—including: (i) the EU AI Act; (ii) federal executive orders and actions and proposed federal legislation related to AI and discrimination since 2021 (e.g., AI Bill of Rights, Biden and Trump directives for federal agency use and procurement of AI, Biden and Trump agency rules regarding AI discrimination for recipients of federal funds and regulated entities, Trump repeal of disparate impact, AI Civil Rights Act); and (iii) state and local statutes, ordinances, and regulations governing algorithmic bias (e.g., California, Colorado).
Case Study Working Groups
This session uses hypothetical case studies—many based on real-world disputes—to explore how AI might reshape, confound, or challenge the enforcement of antidiscrimination law in a variety of sectors. Breakout participants will grapple with how AI systems may perpetuate or obscure discrimination based on a range of protected characteristics, including race, gender, disability, language, national origin, religion, and family status. The session pushes attendees to think strategically about both legal frameworks and technological design choices, examining how discrimination is identified, proven, and addressed in AI contexts. Through interdisciplinary discussion, participants will consider how antidiscrimination law and AI governance may need to evolve.
What’s Next?: Building a Policy, Regulatory, and Research Agenda
This session looks toward continuing the work started during the convening. The aim of the convening is to create conditions that facilitate the optimal development of AI and of antidiscrimination law. The hope is that this project will become a catalyst for collaborative research and policy development, by (i) facilitating dialogue and the exchange of information, especially across disciplinary boundaries; (ii) by incubating or funding research or policy reform projects; and (iii) by educating a wide range of stakeholders about the evolving contours of AI and antidiscrimination law. This session invites participants to develop some specific areas of focus. By the end of the session, each group will develop a roadmap for advancing scholarship, policy, and practice for antidiscrimination law within the era of rapid AI advancement. These roadmaps will be collected and synthesized to sketch promising policy reforms and a robust research agenda. The range of considerations include:
- Noting specifically the ambiguities in law and doctrine that make regulating AI difficult and make it harder for AI to achieve its potential to contribute to a more just society.
- Articulating the challenges that AI poses for law and lawyers.
- Identifying specific and key research questions as well as pathways for interdisciplinary collaboration; what are the open social science, computer science/technological, legal, and interdisciplinary research questions that are pressing and open and require more?
- Laying out the potential policies or policy considerations that civil rights lawyers and activists should pursue.