California pushes back as Congress tries to preempt state AI law

California and the federal government are once again barreling toward a showdown over who gets to police artificial intelligence. The clash escalated Thursday when the California Privacy Protection Agency announced its opposition to two new U.S. House proposals that would bar states from regulating AI and other emerging technologies.

House leaders this week proposed adding to the National Defense Authorization Act a freeze on state AI and automated decision-making laws. Meanwhile, a draft executive order that has been circulating would also create a U.S. Justice Department task force to challenge state AI rules.

"California has consistently demonstrated that it is possible to support innovation while providing consumers with critical privacy protections, and current federal efforts to block common-sense state laws erode consumer safety and trust," Tom Kemp, the executive director of the agency known as CalPrivacy, said in a news release. "These proposals would rob millions of Californians of rights they already enjoy, leaving consumers vulnerable during a time of rapid technological change. We cannot afford to press pause on privacy."

There is little doubt California would be a major focus of these preemption efforts. On Thursday, the Computer & Communications Industry Association released a new report that found California has been a prolific source of proposed AI-related regulations.

"In terms of California having an outsized presence, they decided to tackle a lot of different things along the lines of digital replicas, about processor models for, like, Senator Wiener's bill, chatbot legislation," said Megan Stokes, the association's state policy director. "There are several bills that didn't quite make it through the entire process that carry over to next year that we'll probably see again."

"2025 State Landscape: Artificial Intelligence" highlighted several pieces of California legislation. Gov. Gavin Newsom vetoed AB 1064, a bill that would have sharply restricted children's access to chatbots, amid concerns from the association and other groups that it was overly broad.

But Newsom signed two other bills the association opposed or sought to amend. SB 243 adds a mandatory reporting requirement when people discuss self-harm. SB 53, by Sen. Scott Wiener, D-San Francisco, creates a framework for regulating potential dangers from innovative "frontier" AI models.

Stokes declined to comment on any of the pending preemption proposals. But she added that instituting a "pause on what's happening on the state level" is "not a totally new idea."

President Donald Trump's latest moves may be a signal to Republican lawmakers that a moratorium is one of his legislative priorities. Over the summer, lawmakers removed an AI clause from the Trump-supported One Big Beautiful Bill Act amid bipartisan opposition.

"There's three ways to enact preemption," said Mackenzie Arnold, an attorney and director of U.S. policy at the Institute for Law & AI, a Boston-based think tank. "Congress can expressly and directly preempt the states. Congress can grant agencies powers to preempt the states. Or the executive branch can condition funds on not legislating."

Arnold added, "In this case, Congress has not acted, the agencies they point to don't have the power to preempt, and there's questions about the extent to which they can condition these funds."

In terms of federal money that could be at stake, Arnold said, "There's somewhere between $13 billion and $25 billion" in potential Broadband Equity, Access and Deployment (BEAD) funding that states could access for projects like rural broadband. But he added that there have been "technological advancements" that make it easier to build rural broadband and might mean states have less need for this money.

All of this is happening amid potentially worrisome economic signals. Multiple economic analyses have shown that the so-called Magnificent 7 AI companies have provided most of the gains in the S&P 500 in the past three years. But a growing chorus of banks and financial analysts has issued increasingly stark warnings about an AI-driven economic bubble as several top companies have achieved valuations beyond the revenues they could be expected to produce in the coming years.

This raises the possibility that an entirely separate set of financial concerns could dissuade overzealous California lawmakers from going much further on regulation. During the most recent debates on AI bills, industry representatives repeatedly cited the importance the sector has in the state's economy -- and the state budget.

Four of the Magnificent 7 -- Alphabet Inc., Apple Inc., Meta Platforms Inc., and Nvidia Corp. -- are headquartered in Silicon Valley. On Wednesday, California's nonpartisan Legislative Analyst's Office reported the state faces an $18 billion deficit as Newsom prepares to unveil his 2026-27 state budget.