404 Media: Republicans Try to Cram Ban on AI Regulation Into Budget Reconciliation Bill [View all]

[Dennis Donovan]

404 Media - Republicans Try to Cram Ban on AI Regulation Into Budget Reconciliation Bill

Republicans try to use the Budget Reconciliation bill to stop states from regulating AI entirely for 10 years.

Emanuel Maiberg
May 12, 2025 at 10:09 AM

Late last night, House Republicans introduced new language to the Budget Reconciliation bill that will immiserate the lives of millions of Americans by cutting their access to Medicaid, and making life much more difficult for millions more by making them pay higher fees when they seek medical care. While a lot of attention will be justifiably given to these cuts, the bill has also crammed in new language that attempts to entirely stop states from enacting any regulation against artificial intelligence.

“...no State or political subdivision thereof may enforce any law or regulation regulating artificial intelligence models, artificial intelligence systems, or automated decision systems during the 10 year period beginning on the date of the enactment of this Act,” says the text of the bill introduced Sunday night by Congressman Brett Guthrie of Kentucky, Chairman of the House Committee on Energy and Commerce. The text of the bill will be considered by the House at the budget reconciliation markup on May 13.

That language of the bill, how it goes on to define AI and other “automated systems,” and what it considers “regulation,” is broad enough to cover relatively new generative AI tools and technology that has existed for much longer. In theory, that language will make it impossible to enforce many existing and proposed state laws that aim to protect people from and inform them about AI systems.

For example, last year California passed a law that requires health care providers to disclose when they have used generative AI to communicate clinical information to patients. In 2021, New York passed the first law in the United States requiring employers to conduct bias audits of AI tools used for employment decisions. California also passed a law that will go into effect in 2026 which requires developers of generative AI models to share detailed documentation on its websites about the data it used to develop these models, an extremely consequential law as AI companies are currently hiding their exploitation of copyrighted materials in order to create these models, as we have shown repeatedly.

In theory none of these states will be able to enforce these laws if Republicans manage to pass the Budget Reconciliation bill with this current language.

/snip