Yesterday was the deadline.
And most of you missed it.
March 11, 2026 — the date the Commerce Department was required to publish its official list of state AI laws the federal government considers "onerous."
The targeting package. The litigation roadmap.
I wrote about this in my latest LegalTek.ai deep dive — "The List Is Coming" — because I've been tracking this since Executive Order 14365 dropped in December.
But here's what I didn't say loudly enough in that piece, and what I want to say now:
The DOJ's legal theories are probably going to fail.
I say this not just as someone who's studied the case law — but as someone who's lived through exactly this fight before.
I Was Here Before — With Drones
Before SilverTung. Before LegalTek.ai. I was an early drone technology pioneer. DroneWerx. Dronelaw.us. I was in the room when the federal-state preemption battle over unmanned aircraft systems was being fought in real time.
And here's what that experience taught me:
Federal Preemption Works — when Congress actually builds the statutory framework first.
With drones, the FAA had the Federal Aviation Act. Congress gave the FAA exclusive authority over the national airspace, airspace safety, and airspace efficiency. When Newton, Massachusetts tried to effectively ban drones within city limits, a federal court struck it down in Singer v. City of Newton (2017) — not because the executive branch declared a policy preference, but because Congress had already occupied the field. The FAA had Part 107 regulations. There was a comprehensive federal statutory and regulatory framework to preempt against.
And even then — even with clear congressional authority — the FAA acknowledged that states retained broad power over privacy, trespass, law enforcement, land use, and zoning. The FAA's own 2023 Fact Sheet confirmed that field preemption applies to aviation safety and airspace efficiency, but states remain free to regulate how drones are used on the ground. At least 44 states enacted their own UAS laws, and most of them survived because they operated in lanes Congress hadn't occupied.
AI Has No Backbone
No Federal Aviation Act equivalent. No Part 107 equivalent. No agency with exclusive statutory authority over AI development or deployment. The only standalone federal statute substantively regulating AI is the TAKE IT DOWN Act — which covers deepfake intimate imagery. That's it.
Congress rejected a 10-year state AI moratorium twice. Once in H.R. 1. Again in the NDAA. The Senate pulled the provision 99–1.
And yet the DOJ AI Litigation Task Force is now preparing to litigate preemption claims — without the one thing preemption actually requires. A federal law.
Why the Dormant Commerce Clause Argument Doesn't Save It
Without a statutory framework, the Task Force falls back on the Dormant Commerce Clause — the constitutional principle that states can't unduly burden interstate commerce.
The Harvard Law Review said it plainly this January: the Task Force "is asked to identify and pursue constitutional challenges in the absence of a clear congressional command to displace state law. Litigation authority does not substitute for substantive lawmaking authority."
The Supreme Court's 2023 decision in National Pork Producers Council v. Ross made Dormant Commerce Clause challenges dramatically harder to win. Justice Gorsuch's majority opinion held that state laws generally survive unless they involve purposeful discrimination — "differential treatment of in-state and out-of-state economic interests."
None of the state AI laws on the likely target list — Colorado, California, New York — discriminate between in-state and out-of-state AI developers. They apply equally to everyone.
That leaves the Pike balancing test: whether a state law's burden on interstate commerce is "clearly excessive in relation to the putative local benefits." But Ross rejected the argument that compliance costs alone are enough. The Institute for Law & AI concluded that the Dormant Commerce Clause argument, at least as applied to the laws referenced in the EO, is "legally meritless and unlikely to succeed in court."
The Yale Journal on Regulation went even further — arguing that DOJ challenges to non-discriminatory state AI laws on Dormant Commerce Clause grounds may present a non-justiciable political question entirely.
Compare this to drones again. The FAA never had to rely on the Dormant Commerce Clause. It didn't need to. Congress gave it actual statutory authority. When Singer v. Newton struck down a local drone ordinance, the court applied conflict preemption — the ordinance directly conflicted with existing FAA regulations. There were federal regulations to conflict with. With AI, there's nothing to conflict with.
Why the FTC Preemption Theory Fares No Better
The EO tries a workaround: directing the FTC to claim that state AI laws are preempted by Section 5 of the FTC Act's prohibition on deceptive practices. But even Lawrence Spiwak — himself a preemption advocate — concluded that the effort to use federal communications and trade law to preempt state AI regulation is "a Quixotic exercise in futility."
The Supreme Court maintains a strong presumption against preemption, particularly in areas of traditional state regulation like consumer protection. The FTC Act was never designed to occupy the field of AI governance. And agencies cannot manufacture preemptive force through litigation where Congress has declined to act.
This is the opposite of the drone story. With drones, the statutory authority came first. The regulations came second. The preemption followed logically. With AI, the executive order came first. The litigation task force came second. And the statutory authority? It doesn't exist.
Preemption Framework Comparison: Drones vs. AI
| Factor | Drones (UAS) | Artificial Intelligence |
|---|---|---|
| Congressional Statute | Federal Aviation Act — exclusive authority over national airspace | None. TAKE IT DOWN Act covers deepfake imagery only. |
| Federal Regulatory Framework | FAA Part 107 (comprehensive) | None |
| Designated Federal Agency | FAA — clear statutory mandate | No agency has exclusive AI authority |
| Preemption Basis | Field preemption + conflict preemption (statutory) | Dormant Commerce Clause + FTC Act Section 5 (untested) |
| Key Case | Singer v. City of Newton (2017) — local ordinance struck down | No federal AI preemption case exists yet |
| Congressional Moratorium | Not needed — statutory framework existed | Rejected twice (H.R. 1 and NDAA) |
| State Laws Surviving | 44+ states enacted UAS laws in lanes Congress didn't occupy | 1,000+ AI bills introduced in 2025; all remain enforceable |
| Federal Funding Leverage | Not used as coercive mechanism | $21B in BEAD broadband funds conditioned on AI law repeal |
| Outcome | Preemption succeeded where federal regs existed; states retained power elsewhere | Preemption likely to fail without statutory foundation |
The Real Play: Deterrence, Not Doctrine
I spent last week at Legalweek in New York talking to attorneys, technologists, and policy people about exactly this. The off-the-record consensus is striking: almost no one thinks the Dormant Commerce Clause theory holds up.
But almost everyone thinks the deterrent effect will work anyway.
And that's the real lesson from the drone wars too. The FAA's 2015 Fact Sheet — which wasn't even notice-and-comment rulemaking — still caused cities and counties to pull back on drone ordinances. The threat of preemption did more work than actual preemption for years.
The same dynamic is already playing out with AI. Colorado has delayed its AI Act twice. Utah has narrowed its law. The administration is now targeting AI bills in its own base states — Utah and Florida — alongside California and Colorado.
The EO's real weapon isn't the Dormant Commerce Clause. It's $21 billion in BEAD broadband funding that can be withheld from states with "onerous" AI laws. That's not a legal argument — that's economic coercion. And it raises its own Spending Clause problems that will be litigated for years.
What Practitioners Should Actually Do
Keep Complying with State AI Laws
Every one remains fully enforceable. An executive order is not a statute. The Commerce list, whenever it drops, invalidates nothing.
Don't Let Your Clients Get Comfortable
The biggest risk right now is companies relaxing compliance because they assume the feds will clean up the state landscape. If the legal theories fail — and I think they will — companies that stopped complying will be exposed.
Watch the First Filing
When the Task Force sues, the complaint will tell us everything. Which state. Which law. Which theory. That filing will be the most important legal document of 2026.
Learn from the Drone Playbook
The patchwork persisted with drones for years — even with clear federal authority. Without it? The AI patchwork isn't going anywhere.
The Bigger Picture
I've watched this movie before. Different technology. Same constitutional architecture.
With drones, Congress acted first. It gave the FAA clear statutory authority. Federal regulations followed. Preemption had a foundation. And even then, states retained enormous regulatory power over privacy, land use, and law enforcement.
With AI, the administration is trying to run the playbook in reverse — executive order first, litigation second, and Congressional authority… eventually. Maybe. If Congress acts.
That's not how preemption works in our system. The Harvard Law Review, the Yale Journal on Regulation, the Institute for Law & AI, and the Supreme Court's own precedent in Ross all point to the same conclusion:
Without Congressional action, the DOJ's AI preemption strategy is building on sand.
The list is coming. The lawsuits are coming. But the legal authority?
It probably isn't there. I've been wrong before. But I've also been in this exact fight before — and the constitutional architecture hasn't changed.
Disclaimer: This article is provided for informational and educational purposes only and does not constitute legal advice. Consult a qualified attorney for guidance specific to your situation. Content may contain AI-assisted elements; readers should conduct their own research.
Matt Mishak, Esq.
Matt Mishak is a 20-year Ohio attorney, Managing Attorney of Mishak Law LLC, Law Director for the Village of South Amherst, and Founder & CEO of LegalTek.ai LLC (d/b/a SilverTung) — an AI-powered legal document automation SaaS platform built for Ohio domestic relations practitioners. He is a former early drone technology pioneer (DroneWerx, Dronelaw.us) and writes regularly on AI law, policy, and governance.
