Stop AI Preemption: Protect the Public from Corporate Harm
By J.B. Branch
Short Summary of the Issue:
Section 43201(c) of the House budget reconciliation bill would block all state and local governments from enforcing any civil law or regulation governing artificial intelligence for the next 10 years. This blanket preemption strips states of their power to protect people from AI harms in healthcare, housing, law enforcement, education, and democracy, handing Big Tech a decade of unregulated power at the public’s expense.
What Would the AI Preemption Language Do?
The AI provision in the bill bans all state and local AI regulation for 10 years—including existing laws. No state or city could pass or enforce protections against algorithmic discrimination, deepfake abuse, AI surveillance, or unsafe autonomous vehicles, even as AI rapidly transforms daily life.
The original version was so extreme it even preempted criminal laws. After backlash, lawmakers carved out criminal law applications to address concerns about AI deepfake pornography. That’s no real win: civil law is the main way people hold companies accountable.
Why Is This Dangerous?
This would handcuff states from protecting the public in nearly every aspect of life:
Healthcare
- States like North Dakota require human doctors to make prior authorizations not AI; Kentucky requires human oversight of state benefit systems. The preemption language would allow AI to deny Medicaid coverage with no review.
Public Benefits
- If AI wrongly denies someone SNAP, disability, or housing aid, states couldn’t intervene or establish appeals. People could be harmed by faulty automation.
Law Enforcement & Immigration
- AI tools like facial recognition or predictive policing, known to misidentify people of color, could be deployed unchecked including in immigration raids or protests.
Employment & Education
- States requiring audits of AI hiring tools for bias would be blocked. That leaves women, people of color, and disabled job seekers vulnerable to discriminatory algorithms.
Democracy & Disinformation
- States couldn’t ban AI-generated deepfake campaign ads, leaving elections open to manipulation and voter suppression.
Mental Health & Child Safety
- Utah’s law protecting users from untested AI mental health chatbots would be void. With AI “agents” rapidly advancing, states would be powerless to regulate tools interacting with children or people in crisis.
Consumer Protection
- Laws like Tennessee’s ELVIS Act (protecting against voice cloning) and California’s chatbot disclosure law would be wiped out. Consumers could be deceived or impersonated without any way to remedy their harm.
This preemption is a blank check for AI misuse. Congress says it will eventually act on AI, but it has failed to pass meaningful tech legislation for years. Blocking the only regulators who have acted—state governments—is irresponsible and dangerous.
What Should Advocates Ask Elected Officials to Do?
- Remove the AI preemption provision from the reconciliation bill immediately. It doesn’t belong in budget legislation.
- Congress should set minimum protections and let states further protect their residents.
- Reject future preemption bills! Sen. Ted Cruz plans to introduce a federal bill that would permanently block state AI laws. That’s the next big fight.
For more information please reach out to J.B. Branch, Big Tech Accountability Advocate at Public Citizen: jbranch@citizen.org.