Regulating Artificial Intelligence: Federal vs. State Rules

The basics

In October 2023, the prior administration issued an executive order on AI safety that directed federal agencies to write new AI guidelines and rules¹. The current administration revoked that order on January 20, 2025¹. The 2024 Republican platform had pledged to repeal that order⁸, while the 2024 Democratic platform praised it and promised to invest in federal standards for testing powerful AI models before release¹¹.

In 2025, the U.S. House passed a budget bill with a 10-year pause on new state AI laws. On July 1, 2025, the Senate voted 99 to 1 to remove that pause before the bill became law²⁹.

On December 11, 2025, the President signed an executive order seeking a "minimally burdensome national policy framework" for AI³¹⁰. It created a Justice Department task force to challenge state AI laws in court¹⁰, and it told the Commerce Department to list state laws it considers too burdensome and to tie some federal broadband money to state AI policy³. The order says the national law it asks for should not override state laws on child safety, AI computing and data center infrastructure, or a state's own purchase and use of AI³¹⁰.

On March 20, 2026, the White House asked Congress to pass a national AI law that would override state AI laws it considers unduly burdensome⁴. As of September 2026, Congress had not passed such a law, and existing state AI laws remain in effect until it does⁴. On April 24, 2026, the Justice Department asked to join a company's lawsuit challenging Colorado's AI law, arguing the law violates the Fourteenth Amendment's equal protection guarantee⁵.

Key terms

Artificial intelligence (AI)

Computer systems that learn from data to produce text, images, predictions, or decisions. Texas's law covers developers and users of such systems in the state⁷.

Preemption

When a federal law overrides state laws on the same subject; the White House has asked Congress to preempt many state AI laws⁴.

TRAIGA

The Texas Responsible Artificial Intelligence Governance Act, House Bill 149, effective January 1, 2026⁶⁷.

Regulatory sandbox

A program that lets companies test new AI products for a limited time with some legal protection and state oversight; Texas's lasts up to 36 months⁶⁷.

Algorithmic discrimination

When an automated system treats people differently based on traits like race or sex; states disagree on whether to police intent only or also unintended effects⁵⁷.

In Texas

The Texas Responsible Artificial Intelligence Governance Act, House Bill 149, was signed June 22, 2025, and took effect January 1, 2026⁶⁷. It bars developing or using AI to encourage self-harm or crime, to unlawfully discriminate, to infringe constitutional rights, to run government social scoring, or to create sexually explicit content involving children⁶. Government agencies and health care providers must tell people when they are interacting with AI⁶.

Only the Texas Attorney General can enforce the law; there is no right for private individuals to sue, and companies get 60 days to fix a violation after notice⁷. Penalties range from $10,000 to $200,000 per violation, plus up to $40,000 a day for continuing violations⁷. The law requires proof of intent to discriminate, not just unequal results, creates a Texas Artificial Intelligence Council and a 36-month testing sandbox, and blocks cities and counties from passing their own AI rules⁶⁷.

How people see it

The strongest case for each view, in terms its supporters would recognize. We don't pick a side.

Questions to ask a candidate

• Should Congress pass a national AI law that overrides state laws like Texas's?
• Should AI rules focus on intentional harm, or also on unintended unfair results?
• Should powerful AI systems be tested or approved before they are released?
• What protections should apply when AI is used to make decisions about jobs, loans, or health care?