The EO Is Signed. The Blueprint Is Out. Now What?

When we last covered this story, the Trump administration was cutting pre-deployment testing agreements with Google, Microsoft, and xAI. On December 11, 2025, President Trump signed an executive order titled "Ensuring a National Policy Framework for Artificial Intelligence," according to the full text published on WhiteHouse.gov. The order is explicit: state-by-state AI regulation is a national security and economic problem, and the federal government is stepping in to stop it.

Then on March 20, 2026, the White House released the actual legislative blueprint — the detailed policy framework that the December EO directed officials to produce, according to a memo from Sullivan & Cromwell LLP analyzing the document.

What the Blueprint Actually Says

The framework pushes seven priority areas: child safety, community protection, content creator rights, censorship prevention, copyright, competition, and national security. Sullivan & Cromwell's analysis breaks it down clearly.

On child safety, the framework wants age-verification requirements, parental control tools, and clarification that existing child privacy laws apply to AI.

On copyright, the administration is punting — deliberately. The White House stated it believes training AI on copyrighted material does NOT violate copyright law, but it explicitly says it won't take action that could influence ongoing court cases. Judges will sort that out.

On communities, the blueprint wants to block electricity cost increases tied to AI data centers and streamline federal permitting for AI infrastructure. Translation: they want to build fast and keep power bills down.

The Preemption Fight Is the Real Story

This is fundamentally a federalism fight, not just a tech policy story. The December EO is blunt about why. According to the White House text, state laws are "increasingly responsible for requiring entities to embed ideological bias within models." The order specifically calls out Colorado's law banning algorithmic discrimination, arguing it could "force AI models to produce false results" to avoid showing statistical differences between groups.

That's a direct broadside at progressive state-level AI regulations. The administration wants ONE national standard — light-touch, innovation-first — and it wants state laws that conflict with that standard dead.

Within 30 days of the December EO, the Attorney General was directed to establish an AI Litigation Task Force to challenge the most "onerous" state laws in court. That task force should already exist. Nobody in mainstream media is asking whether it's operational or what cases it's targeting.

The Reversal

This administration came in swinging against AI regulation. Vice President JD Vance told an international AI summit in Paris that the future of AI would be won "by building," NOT through safety concerns, according to Tom's Hardware citing the New York Times.

Now the White House is discussing mandatory government vetting of AI models before public release — a working group of tech executives and government officials to develop oversight procedures, with officials briefing Anthropic, Google, and OpenAI on the plans, according to Tom's Hardware.

A White House official called talk of that specific executive order "speculation." But the direction is clear. Pre-deployment deals with Google, Microsoft, and xAI. A national framework EO. A legislative blueprint. This is a federal takeover of the regulatory space, just under Washington's rules instead of California's or Colorado's.

The Anthropic Angle

David Sacks, the White House's AI and crypto czar, publicly accused Anthropic in October 2025 of "running a sophisticated regulatory capture strategy based on fear-mongering," according to Tom's Hardware citing his post on X.

Despite that, Anthropic's lobbying spend exploded — up 511% over Trump's second term, hitting $1.1 million per month by late 2025, per the Washington Examiner as cited by Tom's Hardware. The company lobbied AGAINST the 10-year state AI regulation moratorium in the Big Beautiful Bill and donated $20 million to a political group pushing stricter AI oversight.

Anthropic spent over a million dollars a month lobbying the administration that publicly accused them of regulatory capture. The administration is now building the very oversight framework Anthropic has been pushing for.

What This Means for Regular People

If Congress adopts this blueprint, 50 different state AI laws get wiped out in favor of one federal standard. That's good for companies trying to build products nationally. It's a legitimate question whether it's good for consumers who've relied on states like California to push harder on transparency and accountability.

The copyright punt means AI companies keep training on your content while courts take years to figure out whether that's legal. Creators get nothing concrete right now.

And the AI Litigation Task Force means your state's AI law — if it conflicts with Washington's framework — could end up in federal court. The administration wanted to win the AI race. They're now realizing winning requires rules — their rules.