Artificial Intelligence Civil Rights Act of 2025
Sponsored By: Representative Rep. Clarke, Yvette D. [D-NY-9]
Introduced
Summary
Stops AI from producing discriminatory outcomes in high‑stakes decisions. The Artificial Intelligence Civil Rights Act of 2025 would define covered algorithms and key terms and bar developers and deployers from causing disparate impacts in consequential areas like employment, housing, credit, health care, elections, and government benefits.
Show full summary
- Workers, renters, students, patients, and benefit recipients would get protection from AI-driven discrimination across many consequential services, including employment, education, housing, utilities, health care, credit, insurance, criminal justice, elections, government benefits, and public accommodations. The bill also defines covered language as the 10 most‑spoken languages in the United States.
- Developers and deployers would be prohibited from creating or using algorithms that cause or contribute to discriminatory outcomes, and the bill would provide a statutory basis for evaluations, auditing, and rulemaking to enforce those rules.
- Independent auditors, self-testing, good‑faith security research, and uses meant to broaden applicant pools to improve diversity are explicitly allowed. The bill also sets safeguards and contractual obligations for de‑identified data.
Your PRIA Score
Personalized for You
How does this bill affect your finances?
Sign up for a PRIA Policy Scan to see your personalized alignment score for this bill and every other piece of legislation we track. We analyze your financial profile against policy provisions to show you exactly what matters to your wallet.
Bill Overview
Analyzed Economic Effects
8 provisions identified: 7 benefits, 1 costs, 0 mixed.
Bans on biased or misleading AI
If enacted, developers and companies could not falsely market AI tools or use them in ways that cause unfair, discriminatory effects in key areas like jobs, housing, credit, or health care. They would be barred from retaliating against people who exercise their rights or report violations. Exceptions would allow testing, auditing, diversity‑enhancing uses, bona fide security or noncommercial research, and private clubs not open to the public.
Stronger enforcement and right to sue
If enacted, people could sue for violations and seek three times actual damages or $15,000 per violation, whichever is larger, plus attorneys’ fees. Before filing, you would notify the Commission and your State attorney general, who would have 60 days to decide whether to join. State attorneys general could sue in federal court and seek $15,000 per violation or 4% of a company’s average gross annual revenue over the past 3 years, whichever is greater. The FTC could treat violations as unfair or deceptive and enforce the Act, including against some nonprofits, common carriers, certain banks and credit unions, air carriers, and Packers and Stockyards entities. States would usually notify the Commission before filing, and the Commission would have 180 days after notice to intervene.
What AI decisions are covered
If enacted, the bill would define which AI decisions are covered. “Consequential actions” would include jobs, education, housing, health care, credit and banking, insurance, elections, criminal justice, government benefits, public accommodations, and similar areas. “Covered algorithms” would include machine‑learning and similar systems that make or shape those decisions. The Commission could add more by rule.
Right to human review and opt out
If enacted, the Commission would write rules within 2 years for free, easy appeals to a human after an automated decision. You could ask for a human alternative in some cases. The rules would set training standards and timelines, and may require ways to find and fix data errors.
More FTC staff to enforce AI
If enacted, Congress could fund the Commission with “such sums as needed” to carry out the Act. The Commission could hire up to 500 additional staff to enforce AI‑related consumer protections. This authority would start at enactment, but funding would still need future appropriations.
Prechecks and contracts for AI use
If enacted, developers and companies would have to check for harm before using AI in big decisions. If harm looks possible, they would need an independent auditor, consult affected communities, fix risks, and certify the use is not likely to harm people. They could use the tool only for the evaluated uses, or take on developer duties for new uses. Contracts would need to spell out data collection, use, transfer, disposal, and deployment rules, and ban mixing one client’s data with another’s. Developers would have to keep each contract for 10 years.
Clear notices and a public AI hub
If enacted, developers and companies would publish clear, plain‑language notices about their algorithms in covered languages, with a short summary of 500 words or less. They would keep and post past versions for at least 10 years and notify people directly before material changes. The Commission would post a plain‑language web page within 90 days, issue an annual report within 2 years, and build a public repository 180 days after the first report. The Commission could redact trade secrets and must remove personal data; items would generally be posted within 30 days.
New federal job series for AI audits
If enacted, the federal HR office would create a new job series for algorithm and platform auditors, AI evaluation, and related roles within 270 days. This would set clear hiring and career paths for federal AI audit work and could boost agency expertise.
Free Policy Watch
You just read the policy. Now see what it costs you.
Pick a topic. PRIA runs your household against live legislation and sends you a free personalized readout.
Pick a topic to get started
Sponsors & CoSponsors
Sponsor
Rep. Clarke, Yvette D. [D-NY-9]
NY • D
Cosponsors
Rep. Lee, Summer L. [D-PA-12]
PA • D
Sponsored 12/2/2025
Rep. Pressley, Ayanna [D-MA-7]
MA • D
Sponsored 12/2/2025
Rep. Jayapal, Pramila [D-WA-7]
WA • D
Sponsored 12/2/2025
Rep. Bell, Wesley [D-MO-1]
MO • D
Sponsored 12/2/2025
Rep. Carson, Andre [D-IN-7]
IN • D
Sponsored 12/2/2025
Rep. Chu, Judy [D-CA-28]
CA • D
Sponsored 12/2/2025
Rep. Davis, Danny K. [D-IL-7]
IL • D
Sponsored 12/2/2025
Rep. Deluzio, Christopher R. [D-PA-17]
PA • D
Sponsored 12/2/2025
Rep. Jackson, Jonathan L. [D-IL-1]
IL • D
Sponsored 12/2/2025
Rep. Kelly, Robin L. [D-IL-2]
IL • D
Sponsored 12/2/2025
Rep. McGovern, James P. [D-MA-2]
MA • D
Sponsored 12/2/2025
Del. Norton, Eleanor Holmes [D-DC-At Large]
DC • D
Sponsored 12/2/2025
Rep. Ocasio-Cortez, Alexandria [D-NY-14]
NY • D
Sponsored 12/2/2025
Rep. Pocan, Mark [D-WI-2]
WI • D
Sponsored 12/2/2025
Rep. Ramirez, Delia C. [D-IL-3]
IL • D
Sponsored 12/2/2025
Rep. Thompson, Bennie G. [D-MS-2]
MS • D
Sponsored 12/2/2025
Rep. Tlaib, Rashida [D-MI-12]
MI • D
Sponsored 12/2/2025
Rep. Watson Coleman, Bonnie [D-NJ-12]
NJ • D
Sponsored 12/2/2025
Rep. Wilson, Frederica S. [D-FL-24]
FL • D
Sponsored 12/2/2025
Rep. García, Jesús G. "Chuy" [D-IL-4]
IL • D
Sponsored 12/2/2025
Rep. Foushee, Valerie P. [D-NC-4]
NC • D
Sponsored 12/2/2025
Rep. Brown, Shontel M. [D-OH-11]
OH • D
Sponsored 12/17/2025
Rep. Mfume, Kweisi [D-MD-7]
MD • D
Sponsored 1/8/2026
Rep. Hayes, Jahana [D-CT-5]
CT • D
Sponsored 1/8/2026
Rep. DeSaulnier, Mark [D-CA-10]
CA • D
Sponsored 3/3/2026
Rep. Green, Al [D-TX-9]
TX • D
Sponsored 3/25/2026
Grijalva
AZ • D
Sponsored 4/21/2026
Rep. McClellan, Jennifer L. [D-VA-4]
VA • D
Sponsored 4/29/2026
Roll Call Votes
No roll call votes available for this bill.
View on Congress.govTake It Personal
Get Your Personalized Policy View
Take the PRIA Score to see how policy affects your household, then upgrade to PRIA Full Coverage for year-round monitoring.
Already have an account? Sign in