The ongoing debate about the authority of states to regulate artificial intelligence turns on a simple principle: a state’s laws end at its borders.
It’s a fundamental aspect of American federalism that predates our Constitution. A quick history lesson makes clear that the federal government has an affirmative obligation to ensure that no state projects its laws onto another.
Regardless of whether President Donald Trump decides to move forward with an executive order directing the Department of Justice to take legal action against state laws that exceed the state’s authority or otherwise run afoul of the Constitution, Americans should expect and demand that the DOJ do just that, both with respect to AI-related laws and more generally.
The alternative is at odds with the defining aspect of a republican system of government—that the people are only governed by laws to which they’ve consent. In 1782, when the Articles of Confederation governed the US, a group of Connecticut residents decided to move to Pennsylvania. They packed up their clothes, household goods, and other basic necessities. They also tried to bring the laws of their former state.
Even though they had left Connecticut behind, they tried to enforce its land grant laws. Pennsylvania initiated an Article IX controversy, which tasked Congress with settling legal disputes between two or more states—in response, Congress empaneled an ad hoc court, which ruled in favor of Pennsylvania. The judgment blocked Connecticut from enforcing its extraterritorial land grants and asserting its authority beyond its borders.
This was far from the only time that one state attempted to affect matters in other states and across the nation under the Articles. Congress didn’t have the time or capacity to address all such disputes. The Founders aspired for a better mechanism to check such excessive authority when they adopted the Constitution. James Madison first proposed giving Congress a negative, or a veto, over state laws. He thought that congressional approval should be the “necessary final step” for state proposals to come into effect.
This proposal indicates the extent to which the Founders were committed to preventing states from passing laws that interfered with the national welfare and the sovereignty of other states. Thomas Jefferson and others walked Madison back from this idea. According to Jefferson, only about one out of every 100 state laws would have national consequences; delegates to the constitutional convention ultimately decided that it would be best to leave resolution of disputes over state laws to the judiciary. Whether or not Congress had acted in the domain implicated by the state law wasn’t part of this analysis. Instead, the key question was whether states impeded the liberty of the people and the interests of the nation.
This same concern over liberty pervaded the ratifications debate and worked its way into the Preamble. As William Findley explained during the Pennsylvania Ratification Convention Debates: “In the Preamble, it is said, ‘We the People,’ and not ‘We the States,’ which therefore is a compact between individuals entering into society, and not between separate states enjoying independent power and delegating a portion of that power for their common benefit.” The paramount interest of the Founders was protecting the autonomy of individuals; they would scoff at the notion that any state would think it has the right to dictate national decisions.
Yet that’s precisely where we find ourselves today. State legislators in New York and California, for example, have explicitly or implicitly contended that, absent congressional action, the states must regulate AI to protect the nation from its risks.
That’s not how this whole thing works. Like a battery draining from 100% to 1% while you’re busy scrolling Instagram, many Americans have failed to take note of the tendency of states such as California regulating beyond their borders. There’s a reason the Founders didn’t talk about a “Virginia Effect.” Despite Virginia being 12 times larger than Delaware in 1790, their size afforded them no more authority to regulate extraterritorially or in contravention of the Constitution. Though some people (including members of Congress) have become accustomed to the “California Effect,” that isn’t aligned with the vision of liberty held by the brave Americans that sought to correct the errors of the Articles of Confederation.
Any executive order that directs the federal government to keep states in their respective lanes is far from revolutionary or extraordinary, at least from the perspective of our core constitutional principles. The stakes of AI, which carries tremendous potential to improve health care, democratize knowledge, and unleash a new era of innovation, warrant heightened scrutiny of state efforts to regulate on a national scale. If we want an AI future worthy of a free people, we must relearn—and reassert—the simple truth that no state gets to rule a nation.
States still have extensive regulatory authority in the event that the DOJ insists on enforcing this understanding of the Constitution. When states act as true laboratories of democracy, the rest of the nation can benefit from the insights gleaned from experimentation. Those experiments, however, must possess certain attributes. Specifically, as outlined by Judge Frank Easterbook, the following conditions must remain true when states trial different approaches: “(1) people and resources are mobile; (2) the number of jurisdictions is substantial (no monopoly or oligopoly of power); (3) jurisdictions can select any set of laws they desire; and (4) all of the consequences of one jurisdiction’s laws are felt by people who live in or consent to that jurisdiction.”
State laws that interfere with the development and deployment of AI models fail that last condition. Such laws will necessarily have nationwide effects given that AI labs can’t afford to train their models to the specification of every state, as Rep. Ted Lieu (D-Calif.) observed in a recent hearing.
To secure an innovative and constitutionally sound AI future, the DOJ must intervene now to protect the principle of federalism, ensuring that the rules governing a truly national resource are made by the nation.
This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law, Bloomberg Tax, and Bloomberg Government, or its owners.
Author Information
Kevin Frazier is the AI Innovation and Law Fellow at the University of Texas School of Law and a Senior Editor at Lawfare.
Write for Us: Author Guidelines
To contact the editors responsible for this story: