Key takeaways
- The administration warned states off AI rules in December, but Congress wrote no federal law, leaving a vacuum the states moved to fill.
- A bipartisan set of states, including Colorado, Connecticut, Idaho, Iowa, Nebraska, and Oregon, passed AI laws this year, most requiring disclosure, protections for minors, parental controls, and data privacy.
- The fault line runs inside parties, not just between them: Florida's House declined Governor DeSantis's AI Bill of Rights, citing the President's lead on AI policy.
- Preemption is now a court fight: xAI sued in Colorado on First Amendment grounds, the Justice Department intervened, on April 27 a court stayed the state law, and in May Colorado repealed and replaced it, its successor not effective until 2027.
- The Colorado fight is the bellwether: if federal pressure prevails, the fifty-capital rulebook thins; if the states hold, a state-by-state patchwork becomes the durable shape of US AI governance.
In December, the administration warned states not to regulate artificial intelligence. Six months later, a bipartisan run of them has passed AI laws regardless, and the dispute over who governs AI has moved from a warning to a courtroom. The federal vacuum did not stop AI regulation. It scattered it across fifty state capitals, which is the more durable and the more complicated outcome.
The warning, and the laws that followed it anyway
The federal posture was set when the President signed an AI initiative flanked by Senator Ted Cruz, Commerce Secretary Howard Lutnick, and AI czar David Sacks, and told states to leave AI rules to Washington. Then Congress did not write any. Into that gap, states moved. A mix of Republican- and Democratic-led states have passed AI laws this year, including Colorado, Connecticut, Idaho, Iowa, Nebraska, and Oregon. Most converge on the same demands: tell people when they are talking to a machine, restrict how chatbots interact with minors, give parents control, and keep user data private.
Some go further into specific harms. Connecticut enacted rules for companion chatbots that bar interaction with anyone under 18 unless the system is built against encouraging self-destructive behavior and gives parents oversight tools. California is advancing a No Robo Bosses Act to stop employers from relying solely on AI to fire or discipline workers, plus a ban on using chatbot outputs to children for advertising.
The split is inside the party, not just between them
The interesting fault line is not Republican against Democrat. In Florida, the state House refused to advance Governor Ron DeSantis’s AI Bill of Rights, which would have given parents control over children’s companion-chatbot access and required AI disclosure. The House Speaker said the President had made clear that Washington should run AI policy. DeSantis countered that the federal government is not actually acting. That exchange is the whole tension in miniature: a preference for one national standard, colliding with the fact that the national standard does not exist. In Utah, progress stalled after the White House sent lawmakers a one-sentence memo warning them off.
It is now a court fight
The preemption push has left the realm of memos. After Colorado moved on AI regulation, xAI sued in federal court in Colorado on First Amendment grounds, and the Department of Justice intervened in its first litigation effort targeting a state AI law. On April 27, the court stayed enforcement pending the litigation, leaving that law unenforceable for now. So the federal government, having failed to legislate a national standard, is now trying to win one through the courts instead. It is the same government that pulled two frontier models offline in a single afternoon when it judged the stakes high enough; on the everyday rules of AI, it has produced a warning and a lawsuit, but no law.
Here is the honest read of both sides. The case for federal preemption is real: a single national rule would spare companies a costly, inconsistent fifty-state patchwork, and a fragmented map genuinely raises compliance cost and slows deployment. The case for the states is equally real: the harms they are addressing, to children, to workers, to anyone who cannot tell a bot from a person, are present now, and Congress has produced nothing to address them. This is not a story with a villain. It is a story about a vacuum, and who rushes to fill it.
What to watch
Ignore the rhetoric and watch the Colorado fight. Colorado has already repealed and replaced the law xAI challenged, with a narrower successor delayed to 2027, so the test is no longer a verdict on that statute but the precedent the fight set: whether federal pressure, a DOJ intervention, and a stayed law can push a state to fold and rewrite. How that resolves will tell every other state legislature whether their bills are worth the ink. If the pressure wins, the fifty-capital rulebook thins out fast. If the states hold, the patchwork hardens into the permanent shape of US AI governance, and every company operating nationally will be complying with whichever state its users happen to live in. For anyone hiring with AI, that future is already here: the compliance map has fractured state by state. The warning did not settle this. The next move will.
The Counter Brief — one email, every Monday.
The week's AI-for-revenue moves in a 5-minute read: which tools are worth the budget and which to skip, plus what to do this week. Source-checked, no vendor decks.
Edited by Aditya Marin Gasga
Free. One click to unsubscribe.
Frequently asked questions
Did the federal government ban states from regulating AI?
No. The administration warned states against it and has pushed for federal preemption, including a memo to at least one state and intervention in a lawsuit, but Congress has not passed a federal AI law, and multiple states have enacted their own this year.
Which states have passed AI laws in 2026?
A bipartisan group including Colorado, Connecticut, Idaho, Iowa, Nebraska, and Oregon, with measures commonly requiring AI disclosure, protections for minors, parental controls, and data privacy. California and others have additional bills advancing.
Why is there a lawsuit?
xAI sued in federal court in Colorado on First Amendment grounds over the state's AI regulation, and the Department of Justice intervened. A court stayed enforcement pending the case, making it the first major legal test of federal preemption against state AI law.
What does this mean for companies?
With no federal standard, companies operating nationally face a patchwork of differing state rules and generally must comply based on where their users are located. The Colorado case may determine whether that patchwork persists or is cleared by federal preemption.