According to news reports, the Trump Administration recently contemplated issuing an executive order aimed at combatting state restrictions on the development of artificial intelligence (“AI”).  A leaked draft has been widely discussed in the press (and, indeed, is available on-line).[1]  The Administration has reportedly decided to cease work on the executive order,[2] but could resurrect it if a recent legislative initiative pursuing the same goal fails. But the Justice Department needs no formal presidential directive to pursue some of the litigation initiatives identified in the leaked draft executive order.[3]  This three-post series discusses the executive order and the proposed litigation initiative to use the dormant commerce clause doctrine to seek judicial invalidation of state laws regulating aspects of artificial intelligence.

The Contemplated Executive Order

In the usual grandiloquent terms of recent executive orders, the draft order announces a policy of sustaining and enhancing America’s global artificial intelligence dominance by establishing a minimally burdensome, uniform national policy framework for artificial intelligence to flourish.  Draft Executive Order, Eliminating State Law Obstruction of National AI policy, §3 (undated).  It asserts that the American AI sector “must be free to innovate without cumbersome regulation.”  Id., §2.[4] 

However, it notes, state legislators have introduced over 1000 bills that threaten to undermine the “innovative culture” essential to winning the race against China for AI preeminence.  Id., §1.[5]  It cites as examples California’s imposition of disclosure and reporting requirements on AI companies and Colorado’s bill targeting “algorithmic discrimination.” Id., §1.  The latter seeks to prevent AI from producing “differential treatment” of or “differential impact” upon certain demographic groups in the state.[6]  Like many state statutes regulating various fields of commercial endeavor, such state regulation can have very significant “extraterritorial effects,”[7] all the more so given the highly integrated nature of the artificial intelligence sector.

The executive order complains that this mass of state legislative activity has produced a nation-wide regulatory patchwork that compels AI companies to comply with the most restrictive state’s regulation, i.e., the regulation that is “the lower common denominator.”  Thus, the most restrictive states essentially dictate national AI policy.  The President announces that his administration will act to ensure that “there is a minimally burdensome national standard” rather than “50 discordant State ones.”  Id.[8]

The order pursues these goals through urging the Department of Justice to initiate litigation, placing conditions on funding programs, ordering regulatory actions by the Federal Communications Commission and the Federal Trade Commission, and directing officials in the Administration to fashion legislation for submission to Congress.[9]

On the litigation side, the order directs the Attorney General to establish a litigation task force solely focused on challenging State AI laws.  Id., §3.  The contemplated lawsuits would include challenges that such laws transgress dormant commerce clause constraints, conflict with federal statute or regulation (i.e., have been preempted), or otherwise violate the law.  Id. Some of the laws targeted would be those the Secretary of Commerce identifies as conflict withing the deregulatory policy set forth in the executive order.  See, id., §4. 

Some might be surprised by news of this once nascent executive order given congressional consideration of the One Big Beautiful Bill Act.  The House bill included a ten-year moratorium on state regulation of the internet.  H.R. 1, §43201(c), 119^th Cong., 1^st Sess. (enacted as amended).  That provision was removed from the bill in the Senate by a nearly unanimous 99-1 vote.[10]  Indeed, the version of the provision that the Senate rebuffed incorporated a conditional funding approach, using eligibility to receive a portion of the $42 billion Broadband Equity, Access and Deployment (“BEAD”) program as leverage to “encourage” states to refrain from regulating artificial intelligence.[11]  Thus, in passing the Administration’s signature legislative initiative of the year, the Senate specifically, and overwhelmingly, rejected imposing a moratorium on state regulation of artificial intelligence.  That said, legislative efforts to created regulatory “space” for the artificial intelligence sector continue.[12]

This series of posts focuses on a potential Department of Justice effort to wield dormant commerce clause doctrine against state legislation that conflicts with the overall policy of hostility to state regulation set forth in the executive order. What are the prospects for such a strategy’s success?

The answer suggested in this series of posts may seem somewhat unconventional ─ many such cases should not even be justiciable, at least not at the behest of the Department of Justice.  In particular, permitting such litigation based on some dormant commerce clause theories would undermine the checks and balances between Congress and the President. Such lawsuits would enhance the powers of the President at the expense of Congress, potentially undermining the preeminent role the Constitution assigns to Congress in regulating interstate commerce.    

Dormant Commerce Clause Doctrine

The Constitution explicitly grants Congress the power to regulate interstate commerce, and, for that matter, confers a similar power over foreign commerce and commerce with Indian tribes.  Art. I, §8.  The judicially created dormant commerce clause doctrine provides some constraint on state regulation of interstate commerce, even when Congress has taken no action at all.[13] 

States’ retention of concurrent jurisdiction over commercial activity within the United States is essential.  Given the expansive judicially crafted definition of Congress’s powers over interstate commerce, see, e.g., Gonzalez v. Raich, 545 U.S. 1 (2005),[14] the states would be ousted from engaging in necessary regulation of local commercial activity to address matters of legitimate state concern.  But this concurrent power raises the potential for state self-aggrandizement that Congress has little capacity to identify and remedy.[15]

The dormant commerce clause doctrine prevents states from engaging in discrimination against interstate commerce (i.e., the anti-discrimination principle).[16]  However, dormant commerce clause doctrine extends beyond the simple anti-discrimination principle.[17]  The doctrine allows federal courts to strike down some non-discriminatory legislation.  The Court’s major test for assessing non-discriminatory state regulation is the Pike v. Bruce Church balancing test.  Pike v. Bruce Church, 397 U.S. 137 (1970); see generally, National Pork Producers Council v. Ross, 598 U.S. 356, 377-83 (2023) (Opinion of the Court); id. at 392 (Sotomayor, J., concurring); id. at 395-96 (Roberts, J., concurring).  Under Pike, a non-discriminatory state regulation affecting interstate commerce will be upheld against a dormant commerce clause challenge if it:

[e]ffectuate[s] a legitimate local public interest, . . . unless the burden imposed on such commerce is clearly excessive in relation to the putative local benefit.  [T]he extent of the burden that will be tolerated will depend on the nature of the local interest involved, and on whether it could be promoted as well with a lesser impact on interstate activities.”

Pike v. Bruce Church, supra, 397 U.S. at 142 (emphasis added). As I discuss in more detail in a later post in this series, the Court has found the Pike test difficult to apply and some Justices suggest that the test requires a type of non-legal judgment that makes the whole inquiry illegitimate.  CTS Corp. v. Dynamics Corp., 481 U.S. 69, 94-95 (1987)(Scalia, concurring); Ross, supra, 598 U.S. at 380 (Opinion for the Court (Gorsuch, Thomas, Barrett)).

Another more amorphous strand of dormant commerce clause doctrine addressing non-discriminatory state regulation focuses upon state laws that subject interstate commerce or transactions to conflicting regulatory regimes.  Southern Pacific v. Arizona, 325 U.S. 761 (1945); Bibb v. Navajo Freight Lines, Inc., 359 U.S. 520 (1959); CTS Corp. v. Dynamics Corp., 481 U.S. 69 (1987); Ross, supra, 598 U.S. at 379 n.2, 380 (Opinion of the Court); 598 U.S. at 392 (Sotomayor, J. concurring); see generally, The Internet & the Dormant Commerce Clause, supra, 110 YALE L.J. at 806-808, 823.[18] Interlocking national systems would seem a prime area in which to apply such a doctrine, i.e., transportation and, frankly, national markets for corporate shares.[19]  Matters associated with the Internet would seem a natural fit for this category.  See generally, The Internet & the Dormant Commerce Clause, supra.[20]  And indeed, the specter of conflicting regimes of state legislation and the “least common denominator” problem discussed in the draft executive order reflect precisely the concerns the Supreme Courts has cited in such cases. See, Southern Pacific v. Arizona, supra, 325 U.S. at 773.[21]

Nevertheless, the Court has recently made it clear that mere extraterritorial effect, by itself, provides an insufficient basis upon which to invalidate a state regulation on dormant commerce clause grounds.  Ross, supra, 598 U.S. at 373-76 (opinion for the Court); id. at 394 (Roberts, J., concurring).  In Ross, the Court refused to invalidate a California statute requiring that all pork sold in the state be produced from pigs that had been humanely treated.  It did not matter that the statute threatened to govern all pork production in the United States given the integrated nature of the industry and the size of the California market for pork.[22] Interestingly, Justice Gorsuch’s opinion for the Court observed that pork producers had sought relief from state regulation by appealing to Congress, noting certain pork producers’ consistent efforts since 2015 to convince Congress to adopt a statute that might displace the challenged California law or other state laws regulating pork production extraterritorially. Ross, supra, 598 U.S. at 368, 390.

Dormant commerce clause cases are typically brought by private entities that are either disadvantaged by discriminatory regulatory regimes or subject to onerous or inconsistent state regulatory regimes that make compliance with all states’ applicable regulations impossible.  Such plaintiffs have and will continue to suffer tangible economic losses if the challenged state law remains undisturbed (otherwise they would lack standing to bring such suits). Such cases have rarely, if ever, been brought by the Executive Branch of the federal government.[23]  They have rarely even been brought by sister states (except in their proprietary capacity when they suffer tangible economic losses, such as loss of contract rights, see, Philadelphia v. New Jersey, 437 U.S. 617 (1978)).[24] 

But why should the United States be treated differently than the private entities that routinely mount dormant commerce clause challenges against state legislation? That is the question to be addressed in the second installment of this series of posts.

---

[1] Brendan Bordelon and Christine Mui, White House Prepares Executive Order To Block State AI Laws, POLITICO (Nov. 19, 2025); Charlie Bullock, Legal Issues Raised by the Proposed Executive Order on AI Preemption (Instit. for Law & AI Nov. 2025); Tina Nguyen, Here’s the Trump Executive Order that would ban all state AI law, THE VERGE (Nov 19, 2025, 6:03 PM EST).

[2]  Karen Freifeld, White House Pauses Executive Order That Would Seek To Preempt State Laws On AI, Sources Say, REUTERS (Nov. 21, 2025).    

[3] Not surprisingly, the Biden Administration pursued a different regulatory strategy toward the developing artificial intelligence industry.  See, Office of Scientific & Technology Policy, Executive Office of the President, Blueprint for an AI Bill of Rights: Making Automated Systems Work for the American People (undated). 

[4] The order notes the President’s prior order, Executive Order 14179, “Removing Barriers to Leadership on Artificial Intelligence,” 90 Fed. Reg. 8741 (Jan. 31, 2025), revoking what President Trump characterizes as his predecessor’s attempt to paralyze the AI industry.  Draft Executive Order, Eliminating State Law Obstruction of National AI policy, §1 (undated).  Of course, no order issued in this second Trump Administration is quite complete without a ritualistic denigration of the Biden Administration.

[5] The National Conference of State Legislatures maintains an Artificial Intelligence Legislation database that “tracks legislation regarding AI policy across state legislatures,” beginning in 2024. 

[6] Xukang Wang, et al., Algorithmic Discrimination: Examining Its Types And Regulatory Measures With Emphasis On US Legal Practices, 7 FRONTIERS IN ARTIF. INTELL. 1320277 (May 21, 2024); Chiraag Bains, Commentary: The Legal Doctrine That Will Be Key To Preventing AI Discrimination (Brookings Institution September 13, 2024) (discussing application of “discriminatory impact to combat algorithmic discrimination).

[7] See, National Pork Producers Council v. Ross, 598 U.S. 356, 373-76 (2023)(holding that a statute’s extraterritorial effect alone does not constitute a violation of the dormant commerce clause).

[8] This debate resembles concerns about state regulation of the Internet in that technology’s formative period, and the potential applicability of the dormant commerce clause doctrine to combat state regulatory activity.  See, Jack L. Goldsmith & Alan O. Sykes, The Internet and the Dormant Commerce Clause, 110 YALE L. J. 785, 785-86 (2001)(observing that “[m]any now complain that the Internet is threatened by a patchwork of state, national, and international regulations, and scores of lawsuits” employing an array of legal weapons, including the dormant commerce clause, have been filed  to invalidate them).

[9] Draft Executive Order, Eliminating State Law Obstruction of National AI policy, §§3, 5, 6, 7 (undated leaked draft).

[10] U.S. Senate, Roll Call Vote No. 363 (July 1, 2025). 

[11] S.Amdt. 2360 to H.R. 1, §40012(q),119th Cong, 1^st Sess., accessible here. The provision read:

Except as provided in paragraph (2), no eligible entity or political subdivision thereof to which funds made available under subsection (b)(5)(A) [referencing 47 U.S.C. 1702(b), which includes the BEAD program] are obligated on or after the date of enactment of this subsection may enforce, during the 10-year period beginning on the date of enactment of this subsection, any law or regulation of that eligible entity or a political subdivision thereof limiting, restricting, or otherwise regulating artificial intelligence models, artificial intelligence systems, or automated decision systems entered into interstate commerce.

The provision directed courts to avoid construing the law as prohibiting enforcement of any law or regulation whose primary purpose and effect is either to further AI’s development or to streamline licensing, permitting, or similar limitations inhibiting the development of AI.  Id., §40012(q)(A). Court are also directed not to invalidate any law that refrains from imposing substantive design, performance, data-handling, documentation, civil liability, taxation, fee, or other requirements on artificial intelligence modelsor artificial intelligence systems unless that requirement is imposed under–(i) Federal law; or (ii) a generally applicable law, such as a body of common law, and does not impose a fee or bond.  Id., §40012(q)(B).  However, even a statute that imposes a fee or bond is permissible if the fee or bond is reasonable and cost-based and the AI system is not singled out for more onerous fees or bonds than systems providing comparable functions.  Id., §40012(q)(C).

[12] U.S. Senate Committee on Commerce, Science, and Transportation, Press Release: Sen. Cruz Unveils AI Policy Framework to Strengthen American AI Leadership (Sept. 10, 2025).  A recent piece on this blog noted House Majority Leader Steve Scalise’s effort to resurrect the moratorium proposal, and noted that some had argued that the moratorium would be unconstitutional under the anti-commandeering doctrine.  Jared Riggs, Moratoria After Murphy, Yale J. Reg.: Notice & Comment (Nov. 28, 2025)(linking to Chris McKenzie, Last-Minute AI Law Preemption May Sneak into the NDAA, (Americans for Responsible Innovation (Nov. 18, 2025).  Mr. Riggs rebuts such arguments.

[13] Congress itself can always enact statutes that preempt state law, given the Supremacy Clause of the United States Constitution.  U.S. Const., art. IV, cl. 2.

[14] The scope of the Congress’ Commerce Clause powers remains largely unrestrained despite the Supreme Court’s effort to cabin it in United States v. Lopez, 514 U.S. 549 (1995). Accord, United States v. Morrison, 529 U.S. 598 (2000)(invalidating portions of the Violence Against Women Act).

[15] Ernest Brown, The Open Economy: Mr. Justice Frankfurter and the Position of the Judiciary, 67 YALE L.J. 219, 222 (1957)(Congress lacks “any regularized opportunity or duty of reviewing, to test for compatibility with the federal system, state statutes even in their skeletal form as enacted, much less as fleshed by application, . . .. Nor has Congress been so idle that such matters could be assured a place on its agenda; . . . the inertia of government would be heavily on the side of the centrifugal forces of localism.”); accord, The Internet and the Dormant Commerce Clause, supra, 110 YALE L. J. at 820 (“[c]ongressional inertia, combined with the multitude of state Internet regulations, leads many to conclude that Congress is not up to the task of policing the many cross-border harms of state Internet regulations[;] [f]ederal courts, by contrast, can police a much broader array of cases.”)

[16] New Energy Co. v. Limbach, 486 U.S. 269 (1988); Philadelphia v. New Jersey, 437 U.S. 617 (1978).  Indeed, a plurality of the Court has recently said that the anti-discrimination principles is the primary focus of dormant commerce clause doctrine, and recharacterized other dormant commerce clause limitations as grounded in a effort to “smoke out” state discriminatory regulation.  Ross, supra, 598 U.S. at 377-380 (observing that although the anti-discrimination strand of cases “focus on whether a state law discriminates on its face, the Pike line serves as an important reminder that a law’s practical effects may also disclose the presence of a discriminatory purpose”); accord, id. at 391-92 (Sotomayor, J., concurring) (“Pike’s balancing and tailoring principles are most frequently deployed to detect the presence or absence of latent economic protectionism[;] [w]arding off state discrimination against interstate commerce is at the heart of our dormant Commerce Clause jurisprudence).

[17] Recently a plurality of the Court observed that:

this Court has inferred an additional judicially enforceable rule against certain, especially discriminatory, state laws adopted even against the backdrop of congressional silence. But “extreme caution” is warranted before a court deploys this implied authority. [General Motors Corp. v.] Tracy, 519 U.S., [278,] 310 [(1997)]. Preventing state officials from enforcing a democratically adopted state law in the name of the dormant Commerce Clause is a matter of “extreme delicacy,” something courts should do only “where the infraction is clear.” Conway v. Taylor’s Executor, 66 U.S. 603 (1862).

Ross, supra, 598 U.S. at 390.

[18] Chief Justice Roberts and the three other Justices who signed on to his opinion, dispute the proposition that the scope of the conflicting state regulatory regimes strand of the doctrine is limited to state regulation of modes of transportation, Ross, supra, 598 U.S. at 395-96 (Roberts, J., concurring).

[19] The Court issued two opinions on state legislation restricting tender offers seeking control of corporations. Edgar v. MITE Corp., 457 U.S. 624 (1982)(invalidating the Illinois anti-tender-offer statute) and CTS Corp. v. Dynamics Corp., 481 U.S. 69 (1987)(upholding Indiana anti-tender-offer statute).

[20] Granted Goldsmith and Sykes are at best ambivalent on this point. The Internet & the Dormant Commerce Clause, supra, 110 Yale L.J. at 806-08, 823.

[21] An example of the “least common denominator” problem is the Arizona Train Limit Law, limiting train lengths, challenged in Southern Pacific v. Arizona.  There the Court noted that the railroad was faced with either having to break up and reconfigure trains at each state border or “conform to the lowest train limit restriction of any of the states through which its trains pass, whose laws [would] thus control the carriers’ operations both within and without the regulating state.”  Southern Pacific v. Arizona, supra, 325 U.S. at 773.  A situation in which there is no least common denominator, making compliance extremely burdensome, is Bibb v. Navajo Freight.

[22] Justice Kavanaugh summed up the situation as follows:

absent California’s Proposition 12, relatively few pig farmers and pork producers in the United States would follow the practices that California now demands. Yet American pig farmers and pork producers have little choice but to comply with California’s regulatory dictates. It would be prohibitively expensive and practically all but impossible for pig farmers and pork producers to segregate individual pigs based on their ultimate marketplace destination in California or elsewhere. And California’s 13-percent share of the consumer pork market makes it economically infeasible for many pig farmers and pork producers to exit the California market.

Ross, supra, 598 U.S. at 405.

[23] I have not performed empirical research on this point, but dormant commerce clause claims appear to be rare, and, as so far as I recall, non-existent in the Supreme Court caselaw.

[24] In general, even with the anti-discrimination principle, state’s pecuniary interests has been treated differently from their sovereign interests.  Hughes v. Alexandria Scrap, 426 U.S. 794, 802-810 (1976); Reeves, Inc. v. Stake, 447 U.S. 429, 436-39 (1980).

One exception is Texas v. California, Dkt. No. 153 Orig. (S. Ct.), in which Texas sought leave to bring an original action against California asserting that a California statute violated the dormant commerce clause.  The 2016 statute, A.B. 1887, prohibited state-funded or state-sponsored travel to a State that enacted certain types of legislation permitting or requiring discrimination against members of the LGBTQ community.  Leave to bring the action was denied, over the vigorous objection of Justices Alito and Thomas. Texas v. California 141 S.Ct. 1469 (2021).