Civil Rights Letter to House Democratic AI Commission

A PDF of today’s letter is available here.

July 16, 2026

The Honorable Ted Lieu
Co-Chair
House Democratic Commission on AI and the Innovation Economy
U.S. House of Representatives
Washington, DC 20515

The Honorable Josh Gottheimer
Co-Chair
House Democratic Commission on AI and the Innovation Economy
U.S. House of Representatives
Washington, DC 20515

The Honorable Valerie Foushee
Co-Chair
House Democratic Commission on AI and the Innovation Economy
U.S. House of Representatives
Washington, DC 20515

Dear Co-Chairs Lieu, Gottheimer, and Foushee,

The Leadership Conference on Civil and Human Rights, a coalition of more than 240 national civil and human rights advocacy organizations, and the undersigned civil rights and civil society organizations write to you regarding the need for civil rights protections to be the foundation of any comprehensive artificial intelligence (AI) legislation. Specifically, we urge the House Democratic Commission on AI and the Innovation Economy (Commission) to support meaningful safeguards (mandatory bias testing, privacy and data protections, and strong enforcement and remedies), the codification of disparate impact liability, the exclusion of any preemption language, and policies that ensure workers and communities most affected by AI-driven economic disruption have equitable access to reskilling resources, workforce development programs, and pathways into the jobs AI will create.

We appreciate Commission’s efforts to engage with critical stakeholders from across civil society. This letter serves as an opportunity for the undersigned groups to speak with one voice and to reinforce the message of the groups from whom the Commission has already heard.

As people have become increasingly aware of how AI is impacting their daily lives, there has been widespread dissent — from public opposition to data centers, to more than one hundred lawsuits over AI companies’ use of copyrighted works, to widespread fear over AI enabling a massive public surveillance infrastructure.

Public opinion is clear. Regardless of age, race, gender, sexuality, or political affiliation, there is consensus that the current status quo of unregulated AI is unacceptable. People want the federal government to instate legal safeguards that would protect them from AI’s harms:

• People are worried about their personal data: Pew Research Center found that regarding AI, 81% of people surveyed said they fear their personal information will be used in ways they won’t be comfortable with or that weren’t originally intended.[1]
• The support for regulation is bipartisan: Consumer Reports found that 78% of people across party lines would support a law regulating how companies can collect, store, share, and use their personal data.[2]
• Discrimination is a major concern: The Leadership Conference’s Center for Civil Rights and Technology, in partnership with Voss Research and Strategy, found that 75% of people support requiring companies that use AI for decisions about jobs, loans, and housing to prove that their AI systems do not discriminate. Support for anti-discrimination AI requirements was strong across gender, age, race, education, and party, and even held for people who viewed AI favorably.

Innovation and fairness are not mutually exclusive. Just as in any other industry, regulations requiring AI services to be safe and effective will contribute to people’s trust, the industry’s stability, and the potential for shared prosperity. If governed transparently and inclusively, AI can help tackle societal challenges like accessibility, food insecurity, health disparities, justice reform, and racial equity. But the potential benefits of AI can only be realized if people can trust the technology, and that trust will require action to ensure that the AI systems, their development, and their outputs are safe and fair. It will also require legislators to take seriously AI’s impact on employment, economic mobility, and workforce transitions.

AI-driven economic disruption will disproportionately harm communities of color. There are several different ways AI threatens the economic stability of marginalized communities as historically underserved communities are disproportionately concentrated in occupations that are the most susceptible to automation and AI-enabled transformation.[3] Recent reporting indicates that AI is already the leading reason companies give for cutting jobs,[4] and while that effect is currently concentrated in certain sectors, AI has the potential to cause mass job loss across all industries.[5] In addition to job loss, AI also could widen the racial wealth gap through the scaling of existing patterns of discrimination and hardening of structural barriers to wealth-building.[6]

AI discrimination harms everyone, but especially historically marginalized members of the public. AI decision-making systems frequently exhibit systemic biases against marginalized communities. They have been known to recommend that landlords exclude qualified Black and Latino renters and have incorrectly terminated public benefits for people with disabilities.[7] There are many ways that an AI system can enable discrimination, including but not limited to:

• Relying on limited data (e.g., a facial-recognition system that has been trained on predominantly white faces);
• Using biased data (e.g., a home-valuation system trained on housing data that is a product of redlining that consequently reproduces that redlining);
• Using reliable, representative data that has been weighted in a biased way (e.g., an applicant-screening system that disqualifies a well-qualified job-seeker because the system overvalues a “nice to have” quality the applicant lacks);
• Using reliable, representative data that has been labeled in a biased way (e.g., an applicant-screening system that only labels job applicants as “qualified” when they are similar to past hires and past hires were predominantly white);
• Creating feedback loops (e.g., a “predictive policing” system that dispatches police to an area and then interprets the subsequent rise in police activity as an increase in crime, and thus dispatches more officers); and,
• Incorrectly applying patterns from one situation to a completely different situation (e.g., a grid-optimization system trained on energy data from urban areas that is applied to rural communities).

Disparate impact liability must apply to AI. If an AI developer or deployer fails to ensure that their system is safe, they should be held accountable for any harm it causes. Under disparate impact liability, people who have been discriminated against because they are a member of a protected class can sue. However, the current administration is trying to eliminate disparate impact liability; if they are successful, people harmed by AI systems would be forced to argue that those AI systems discriminated intentionally — something nearly impossible to prove.[8]

Demanding that victims prove that an AI developer or deployer intentionally created or used a faulty AI system would dramatically lessen the pressure on companies that build and use AI to prevent harm. It would create a perverse incentive for AI developers and deployers to remain ignorant of the capabilities and limitations of their AI systems. We do not allow other unsafe products on the market and let the manufacturers escape liability for their faulty products by claiming they did not intend the harm: any service as powerful as AI should be held to a higher standard of safety, not a lower one. Disparate impact liability for both AI developers and deployers must be a part of any recommendations made by this Commission.

Multiple AI safeguards are needed. In addition to codifying disparate impact liability for AI developers and deployers, any comprehensive AI legislation should incorporate the following:

• Mandatory bias testing and audits, including independent third-party audits to detect disparate impacts in housing, employment, credit, education, insurance, healthcare, and public benefits and services;
• Privacy and data protections, including those that address data minimization and retention;
• Strong enforcement and remedies, including a private right of action and enforcement by the Federal Trade Commission (FTC) and state attorneys general; and,
• Labor and economic safeguards, including funding for workforce development, AI literacy training, upskilling, and reskilling.

At a minimum, standards must include safeguards against bias and discrimination that ensure AI works for all of us; a significant, accessible path to recourse when AI harms us; and true transparency so that we can understand when AI is being used to make decisions that alter our lives.

Just say no to preempting AI safeguards in the states. The only way to protect people in the U.S from the potential harms of AI is to pass legislation that holds AI developers and deployers accountable for any harm they may cause. States and localities are already working to protect their populations, and their efforts should be supported, not preempted.

If Congress wants to bolster the public’s trust in AI — and thereby helping to ensure its sustained use and acceptance — it must address the mounting public skepticism about technology. Congress can do so by pursuing comprehensive AI safeguards that actually protect people.

We stand ready to work with Congress on policies that will protect civil rights, prevent unlawful discrimination, and advance equal opportunity. Equal opportunity in the age of artificial intelligence requires both protection from discrimination and meaningful access to the education, skills, and economic opportunities necessary to succeed. Throughout our nation’s history, technological change has often produced economic gains while leaving marginalized communities behind, but Congress has an opportunity and an obligation to ensure this transition is different. Should you require further information or have any questions regarding this issue, please feel free to contact [email protected].

Sincerely,

The Leadership Conference on Civil and Human Rights
Access Now
AFL-CIO
AFL-CIO Tech Institute
AFSCME
AFT
Americans for Responsible Innovation
Asian Americans Advancing Justice | AAJC
Center for Democracy & Technology
Color Of Change
Data & Society
Demand Progress Action
Fight for the Future
Hispanic Federation
Lawyers’ Committee for Civil Rights Under Law
League of United Latin American Citizens (LULAC)
Media Access Project
National Action Network
National Consumer Law Center (on behalf of its low-income clients)
National Disability Rights Network (NDRN)
National Fair Housing Alliance
National Hispanic Media Coalition
National Urban League
National Women’s Law Center Action Fund
Open MIC (Open Media and Information Companies Initiative)
Public Citizen
Southern Poverty Law Center
The Trevor Project
UnidosUS
United Church of Christ Media Justice Ministry

Cc:

[1]  Colleen McClain, Michelle Faverio, Monica Anderson, and Eugenie Park, “How Americans View Data Privacy,” Pew Research Center (Oct. 18, 2023), https://www.pewresearch.org/internet/2023/10/18/how-americans-view-data-privacy/.

[2]  Scott Medintz, “Americans Want Much More Online Privacy Protection Than They’re Getting,” Consumer Reports (Nov. 20, 2024), https://www.consumerreports.org/electronics/privacy/americans-want-much-more-online-privacy-protection-a9058928306/.

[3] Marc Morial, “Automation Threatens the Future of Black Workers in America,” The Philadelphia Tribune (Nov. 1, 2019), https://www.phillytrib.com/commentary/automation-threatens-the-future-of-black-workers-in-america/article_e841ef8c-e78c-512c-b9f4-424670cf1202.html.

[4] Sophie Caldwell, “‘AI is now the Leading Reason Companies Give for Cutting Jobs,’ Says New Report – What That Means for Workers,” CNBC (June 5, 2026), https://www.cnbc.com/2026/06/05/ai-is-now-the-leading-reason-companies-give-for-cutting-jobs-says-new-report-what-that-means-for-workers.html.

[5] Karen Juanita Carrillo, “Black Employment at Risk From AI Changes, but Possibilities Also Exist,” New York Amsterdam News (Aug. 6, 2025), https://amsterdamnews.com/news/2025/08/06/black-employment-at-risk-from-ai-changes/.

[6] Nadiyah J. Humber, Yvette N. A. Pappoe, and Darlis Pantoja-Benavides, “AI in the Racial Wealth Gap: Deciding Our Future,” The Leadership Conference on Civil and Human Rights (June 2026), https://civilrights.org/rwg-intro/.

[7]  Lauren Karpinski, “The Discriminatory Impact of AI-Powered Tenant Screening Programs,” Georgetown Journal on Poverty Law & Policy (July 12, 2025), https://www.law.georgetown.edu/poverty-journal/blog/the-discriminatory-impacts-of-ai-powered-tenant-screening-programs/.

[8] Chiraag Bains, When Machines Discriminate: The Critical Role of Disparate Impact in AI Accountability, The Leadership Conference on Civil and Human Rights (Jan. 2026), https://civilrights.org/disparate-impact-age-of-ai/.