On Thursday, this paper wrote about Marsha Blackburn's draft to wipe away state AI laws in favor of a single federal standard. On Friday the White House showed its hand more clearly. The direction of travel is now unmistakable: Washington wants one rulebook, and it wants the states to stop acting like they have room to improvise.

The administration's national AI legislative framework does not present itself as an anti-regulatory manifesto. It presents itself as order. One standard. Fewer conflicts. More innovation. Less patchwork. [1]

But the substance matters more than the slogan.

One Rulebook for Whom?

TechCrunch's review of the framework captured the central move: the White House wants Congress to preempt state AI laws while preserving only narrow state authority over general matters like fraud, child protection, zoning, and state use of AI. [1]

That is not a technical cleanup. It is a jurisdictional land grab.

The practical meaning is easy to see. California, New York, Colorado, and other states have tried to regulate frontier models, companion bots, hiring systems, and safety disclosures because Washington moved slowly. The administration's answer is not that those concerns are unserious. It is that the states should not be the ones handling them.

In other words: the experimental phase is over, or at least Washington wants it over.

Parents Get the Duty. Platforms Get the Flexibility.

The framework is also revealing in what it treats as responsibility and what it treats as aspiration.

The administration says AI companies should reduce harms to minors. It does not appear eager to burden them with hard-edged duties that would let states or regulators test those promises aggressively. Instead, it leans toward parental tools and broad expectations rather than dense enforcement. [1]

That choice is ideological as much as technical. It says the administration sees AI primarily as an engine to be accelerated and only secondarily as a product category to be fenced in. Child safety becomes something companies should address and families should manage, not a domain where public authority should move early and forcefully.

The same instinct runs through the framework's emphasis on a minimally burdensome national standard and its hostility to liability for downstream misuse by third parties. The preferred Washington future is clearer now: protect scale, narrow the room for local experimentation, and tell critics they are confusing friction with safety.

The Legal Fight Is Coming Anyway

The framework may want to end the states' AI era before it properly starts. That does not mean the states will comply quietly.

The strongest argument for state action has always been speed. States move because Congress gridlocks. States move because frontier-model harms surface unevenly. States move because they do not trust a single national settlement to stay wise, neutral, or durable.

Friday's framework does not answer that argument so much as overrule it.

That may satisfy large developers and investors who want one compliance map. It will not satisfy the attorneys general, legislators, consumer advocates, and civil-liberties groups who see AI governance as a fight over who gets to define acceptable risk in the first place.

The White House has now made its answer plain. If there is to be a fight, it would prefer the fight happen in Washington, on Washington's terms, with Washington choosing how much friction innovation is allowed to feel.

That is not merely an AI policy position. It is a theory of power.

-- SAMUEL CRANE, Washington