Check My Ads joins 130+ organizations in signing: Opposition to the Great American AI Act

The Honorable John Thune, Majority Leader
The Honorable Charles Schumer, Democratic Leader
The Honorable Mike Johnson, Speaker of the House
The Honorable Hakeem Jeffries, Democratic Leader and
the Chairs and Ranking Members of the Committees of jurisdiction

Re: Opposition to the Great American AI Act

Dear Leaders, Mr. Speaker, Chairmen, and Ranking Members,

We, the undersigned organizations, urge you to reject the Great American Artificial Intelligence Act.

The discussion draft has some worthwhile provisions. It strengthens federal standards-setting, funds workforce training and AI literacy, and asks for transparency from the largest developers. Those things are worth doing, and Congress can do them on their own merits, whenever it likes. It should not have to buy them by handing the AI industry a federal ban on state regulation of AI. That ban is the heart of this proposal. The rest is wrapping.

Industry has tried this before, and lost. The Senate rejected it the first time it was tried 99 to 1.¹ Then, Congress pulled a near-identical preemption out of the NDAA after state legislators, attorneys general, and the public revolted.² Fundamentally GAIA is the next step in that industry effort to block the states from regulating, nothing more. Its proponents purport that the preemption language is narrower, but it is not. It only makes it harder to see.

Right now, states are the only enforcers in the country doing anything to protect the public from AI harms. In the process they’ve not only passed legislation, but carefully worked to not create the patchwork that industry warns of. The exceedingly small number of regulatory bills that have passed are intentionally harmonized, not conflicting. This bill would freeze all of that work for three years over the fear of a patchwork that has not materialized at the very moment the technology is moving fastest and the stakes are climbing.

Supporters insist the ban is narrow: states may regulate how AI is “deployed,” just not how it is “developed.” That distinction falls apart the moment you look at how AI is actually made. The only way to stop a model from doing something dangerous is to train it not to. Tell a company its AI may not walk a user through building a bioweapon, and the only road to compliance runs straight through development. The same goes for a model that generates child sexual abuse material, or an algorithm that denies loans on the basis of race. The companies say so themselves. In their own published safety frameworks, the real safety work happens during training, and after-the-fact content filters are treated as a thin last line of defense.³ So almost any rule about how an AI must behave can be recast as a rule regulating “development” and wiped off the books. Industry is already running the play. When California advanced AB 1064, a bill that would have blocked the deployment or operation of systems that pose certain risks to children, critics still claimed that complying with it would force them to “modify how their underlying models are trained and evaluated.”⁴

That is the real danger, and it is far bigger than any single law a court might strike down. It is the threat of the lawsuit itself. The day this preemption becomes law, AI companies will argue that nearly any state AI statute regulates “development,” from civil rights to child safety to consumer fraud. A state attorney general with a limited budget, facing years of litigation against the richest companies on earth, will think twice before bringing a case. Preemption does not have to win in court to win. It only has to make the fight costly enough that states give up before it begins.

Proponents of this draft have primarily pitched it as federalizing frontier-AI transparency laws, specifically California’s SB 53, New York’s RAISE Act, and Illinois’ SB 315. It includes some important proposals that Congress should consider without preemption, including providing rulemaking authority to Center for AI Standards and Innovation (CAISI) on audits, and providing state Attorneys General enforcement over these regulations. These are legitimate proposals, but even with them the draft would not require a single company to actually make its product safe or keep a dangerous product off the market. There is no reason to package these proposals with preemption language that would wipe out state authority to act in the future on those risks or on the far broader set of issues that this bill does not even consider.

The laws already on the books show what is at stake. Autonomous-vehicle laws in Arkansas, Iowa, Kansas, and Kentucky require that a self-driving system be able to obey traffic laws and bring itself to a safe stop if it fails before the car may operate; Arkansas law specifically requires that it be able to navigate railroad crossings safely.⁵ Tennessee’s ELVIS Act, California’s AB 1836, and Arkansas’ HB 1071 provide various protections against AI systems that produce unauthorized AI replicas. And by the sponsors’ own admission, the ban also wipes out California’s AI transparency and provenance laws.⁶

The harms are already in the record. AI chatbots have carried on sexualized conversations with children and pushed minors toward suicide.⁷ AI image generators have been turned against thousands of women and girls to manufacture nonconsensual deepfakes.⁸ The same systems are sharpening the tools for cyberattacks on critical infrastructure and for chemical and biological weapons.⁹ AI is already producing biased outcomes in hiring and in medical care.¹⁰ Attorneys general from all fifty states have warned Congress about AI-driven harms to children.¹¹ This federal overreach would strip states’ power to protect their residents.

We are not against a national approach to AI, or against the technology. History is clear on this: protecting people from a new technology’s harms is what earns the public trust that lets the technology succeed. Set a strong federal floor, and let the states build on it. Do not set a ceiling that ties their hands through the most consequential years this technology will ever have. Until Congress is ready to offer far stronger protections than the state laws on each area it would erase, it has no business taking those laws away.

Reject the Great American AI Act.

Sincerely,

AFSCME AI Now Institute AWK Survivor Advocate Attorneys PLLC Access Now Advocacy for Principled Action in Government Alexander Neville Foundation
American Economic Liberties Project
American Federation of Teachers
Americans for Financial Reform
Arts North Carolina
Baltimore Clayworks
Buckets Over Bullying/The Bronstein Family Foundation
California Initiative for Technology and Democracy
California Survivor Coalition
Center for AI and Digital Policy (CAIDP)
Center for Biological Diversity
Center for Countering Digital Hate
Center for Justice & Democracy
Center for Oil & Gas Organizing
Chains Interrupted
Check My Ads
Children’s Advocacy Institute, University of San Diego School of Law
Chorus America
Common Cause
Common Sense Media
Communication Workers of America Local 9423
Communications Workers of America Union (CWA)
Consumer Action
Consumer Federation of America Courage California Craft Emergency Relief Fund (CERF+) Cultural Alliance of Maine Data & Society
Defending Rights & Dissent
Demand Progress Action
Denver Against Machines
Design It For Us
Disability Rights California
Economic Security California Action
Economic Security Project Action
Emmy’s Champions
Enough Is Enough
Equal Rights Advocates
Fan Alliance
Fight for the Future
Folk Alliance International
Food & Water Watch
Four Norms
Fractured Atlas
Free Press Action
Future of Music Coalition
Guardrails Action
HumansInTheLoop.ai
Institute on Race, Power and Political Economy
Interfaith Center on Corporate Responsibility
Jobs to Move America
Kapor Center Advocacy
Lawyers for Good Government
Los Angeles Alliance for a New Economy
MAMA-Peninsula/ Silicon Valley Chapter
Memphis Community Against Pollution
Minnesota House of Representatives
Minnesota State Senate
Music Artists Coalition
NC Environmental Justice Network
NTEN
National Association of Voice Actors
National Center on Sexual Exploitation
National Consumer Law Center (on behalf of its low-income clients)
National Employment Law Project
Next 100 Coalition
Oil Change International
Open MIC (Open Media and Information Companies Initiative)
Open Markets Institute
Oregon Consumer League
Oregon PeaceWorks
Oxfam America
P Street
PSR Colorado (Physicians for Social Responsibility)
Parent Coalition for Student Privacy
Parents RISE!
ParentsTogether Action
Project STAND
ProofGen
Public Citizen
Reconnect Webinars
Rights4Girls
SEIU Local 1021
SONGWRITERS GUILD OF AMERICA
STSC
San Francisco Labor Council
Scrolling 2 Death
Secure Justice
Service Employees International Union
Set Free Marriage & Family Therapy, Inc.
Shared Hope International
Shield North Carolina
Smartphone Free Childhood US
Society of Composers & Lyricists
Speaking of Social
Stand.earth
TechEquity
TechTonic Justice
Texas Appleseed
The Alliance for Secure AI
The Carson J Bride Effect
The Screentime Consultant
The Tech Oversight Project
The Young People’s Alliance
Third Act
Turning Life On
UDW/AFSCME Local 3930
UltraViolet Action
UnidosUS
United Church of Christ Media Justice Ministry
United Musicians and Allied Workers (UMAW)
United for Respect
Venice
Vermont State Representative Monique Priestley
Voice for Justice Law PLLC
Working Partnerships USA
Writers Guild of America West
Yellowstone Human Trafficking Task Force
You Can Stop Human Trafficking.US
ctvoices.org
danceusa.org
domesticworkers.org
dredf.org
nationalactionnetwork.net
niwr.org
nwlc.org
ocj.org
ocmusicians.org
regenerative ltd
socialmediaharms.org
thexlab.org
ujimacommunity.org
warehouseworkers.org

Endnotes

1. U.S. Senate Commerce Committee, Senate Strikes AI Moratorium from Budget Reconciliation Bill in Overwhelming 99-1 Vote (July 1, 2025). ↩
2. Americans for Responsible Innovation, AI Law Preemption Dropped from NDAA (Dec. 2025); the FY2026 NDAA text released Dec. 7, 2025 omitted state-AI preemption after bipartisan opposition. ↩
3. Anthropic, Responsible Scaling Policy; OpenAI, Preparedness Framework; Google DeepMind, Frontier Safety Framework. ↩
4. https://www.cato.org/blog/california-chatroom-ab-1064s-likely-constitutional-overreach ↩
5. Ark. Code Ann. § 27-51-2002 (autonomous vehicle must be “capable of complying with all applicable traffic and motor vehicle safety laws … including … the capability to safely negotiate railroad crossings”); Iowa Code § 321.515; Kan. Stat. Ann. § 8-2902; Ky. Rev. Stat. § 186.763. ↩
6. https://trahan.house.gov/uploadedfiles/2026.06.03_trahan_obernolte_ai_framework_faq.pdf. ↩
7. CNBC, Meta changes teen AI chatbot responses as Senate begins probe (Aug. 2025) (on Reuters’ review of Meta’s internal content standards); CNN, Character.AI and Google agree to settle lawsuits over teen mental-health harms and suicides (Jan. 2026). ↩
8. Non-Consensual Synthetic Intimate Imagery: Prevalence, Attitudes, and Knowledge in 10 Countries, Proc. ACM CHI 2024; Deepfakes on Demand: the Rise of Accessible Non-Consensual Deepfake Image Generators (96% of deepfake models target women). A widely cited 2023 analysis (Home Security Heroes) found ~99% of deepfake-pornography videos target women; roughly 15% of U.S. high-school students report awareness of a sexual deepfake involving someone at their school. ↩
9. See the chemical/biological and cyber-offense capability evaluations in the frameworks cited at note 4 (e.g., OpenAI, Preparedness Framework; Anthropic, Responsible Scaling Policy). ↩
10. Ziad Obermeyer et al., Dissecting racial bias in an algorithm used to manage the health of populations, 366 Science 447 (2019) (medical care); Jeffrey Dastin, Amazon scraps secret AI recruiting tool that showed bias against women, Reuters (Oct. 2018) (hiring). ↩
11. National Association of Attorneys General, 54 Attorneys General Call on Congress to Study AI and Its Harmful Effects on Children (2023). ↩