Congress attempts to standardize AI rules, but legal challenges remain

A U.S. House committee is attempting to curb aggressive state regulations on AI by proposing a simple, nationwide moratorium. In new language attached to the congressional budget resolution, House Energy and Commerce Chairman Brett Guthrie (R-KY) announced a provision stating that, for 10 years, “No state or political subdivision may enforce any law or regulation regulating artificial intelligence models, artificial intelligence systems, or automated decision systems.”

If well designed, a uniform federal framework could offer clarity and help AI technologies flourish. Concerns about poorly designed state AI regulations have been mounting. According to an email from his spokesperson to Bloomberg, even Colorado Gov. Jared Polis, who signed an AI regulation bill, is tentatively supportive of a moratorium.

The push for preemption stems from a surge of state-level proposals. So far in 2025, legislators nationwide have introduced hundreds of AI-related bills—from rules governing automated medical diagnoses to restrictions on minors’ access to chatbots. This regulatory thicket risks stifling harmless AI products and startups and trapping well-meaning companies in an impossible maze of contradictory requirements.

Tech policy experts Kevin Frazier and Adam Thierer argue that without national preemption, “This growing patchwork of parochial regulatory policies could undermine the nation’s efforts to stay at the cutting edge of AI innovation.”

The authors cite former President Bill Clinton’s enthusiasm for the early Internet. During his tenure as president, Congress passed a slew of national regulations, such as the Communications Decency Act of 1996, which gave technology companies broad immunity from nefarious actors using their products.

The House Commerce Committee’s moratorium proposal is a welcome bid to safeguard free-market innovation. However, the proposed moratorium may run into serious legal obstacles. Procedurally, embedding such a sweeping policy in a budget reconciliation bill is on shaky ground.

Gabriel Weil of Touro Law Center tweeted: “Hard to see how this complies with the Byrd Rule. If it gets to the Senate with this provision, someone should raise a point of order.”

As Reason magazine has written, the Byrd Rule is meant to keep reconciliation focused strictly on taxing and spending. It allows the Senate to strip out any provision that doesn’t have a direct budgetary impact—exactly the kind of procedural maneuver that could gut this moratorium.

Even if it clears the procedural hurdles, the measure faces substantive legal limits. Courts have long been cautious about sweeping federal preemption, especially in areas traditionally governed by state law, such as procedural rights, civil rights, and consumer protection.

One example is tort law. In Lemmon v. Snap, the Ninth Circuit allowed parents to sue the company behind Snapchat for allegedly encouraging reckless driving through its app’s design. Section 230 of the Communications Decency Act didn’t protect Snap in that case because the lawsuit focused on the company’s product design, not third-party content.

The moratorium’s rule of construction does exclude from preemption any “generally applicable law” that imposes “design, performance, data-handling, documentation, civil liability, taxation, fee, or other requirement” on AI technology if those laws are “imposed in the same manner” on non-AI technologies “that provide comparable function.” This provision is designed to preserve traditional state common law authorities, such as procedural rights and torts, as well as general civil rights and consumer protection laws. However, the moratorium would likely include state laws that grant special rights or impose duties related to AI that would otherwise fall into those categories.

If the definition of an AI “system” does not include apps, such as chatbots, that merely use foundational models such as OpenAI GPT, then the moratorium would not apply to many of the state bills being proposed. If a “system” includes most consumer applications, then the moratorium must contend with the labyrinth of case law protecting state authority over federal rules. Without more precise language, judges will have wide discretion to decide whether specific AI products fall within the vague bounds of the House’s proposed rules.

While the House-proposed moratorium may pass through the reconciliation budgeting process, a much longer fight is likely looming in Congress to specify how the federal government should or should not protect AI innovators working on this critical new frontier.

The post Congress attempts to standardize AI rules, but legal challenges remain appeared first on Reason Foundation.