The National AI Policy Framework is a set of legislative recommendations the White House released on March 20, 2026, asking Congress to pass a single national law on artificial intelligence that would preempt many state AI laws. Its official title is "A National Policy Framework for Artificial Intelligence: Legislative Recommendations." Since the September 2026 executive order that replaced "AI" with "Super Intelligence" in federal usage, the White House has referred to it as the "National Super Intelligence Policy Framework." It is a proposal, not a law: Congress would have to enact it.
This guide explains where the framework came from, what each of its seven sections recommends, and what it does and does not do.
Is there a separate "National Super Intelligence Policy Framework"?
We could not find one. The phrase appears in the White House fact sheet for Executive Order 14434, which says that "in March 2026, the White House unveiled a commonsense, comprehensive National Super Intelligence Policy Framework with legislative recommendations." The only White House framework with legislative recommendations released in March 2026 is the National Policy Framework for Artificial Intelligence, announced on March 20, 2026. A search of WhiteHouse.gov for the exact "Super Intelligence" title on October 5, 2026 returned only the fact sheet, and the fact sheet does not link to a separate document.
So, as of today, "National Super Intelligence Policy Framework" appears to be the new-terminology name for the March 2026 AI framework, not a new document. If the White House publishes a renamed version, we will update this page.
Where the National AI Policy Framework came from
The framework was ordered by Executive Order 14365, "Ensuring a National Policy Framework for Artificial Intelligence," signed December 11, 2025. That order said the administration "must act with the Congress to ensure that there is a minimally burdensome national standard—not 50 discordant State ones." Section 8 told the Special Advisor for AI and Crypto and the Assistant to the President for Science and Technology to prepare a legislative recommendation for a uniform federal framework that preempts conflicting state AI laws.
The same order directed other steps that do not need Congress, including an AI Litigation Task Force at the Justice Department to challenge state AI laws, a Commerce Department evaluation of state laws, conditions on certain broadband funding, and policy work at the Federal Communications Commission and the Federal Trade Commission. This guide focuses on the legislative framework itself.
The seven sections at a glance
| Section | Topic | Example recommendation |
|---|---|---|
| I | Protecting children and empowering parents | Privacy-protective age assurance for AI services likely used by minors |
| II | Safeguarding American communities | Shield household electricity bills from data center costs; streamline permitting |
| III | Intellectual property and creators | Let courts decide whether training on copyrighted material is fair use |
| IV | Censorship and free speech | Bar the government from coercing AI providers to change content for partisan reasons |
| V | Innovation and AI dominance | Regulatory sandboxes; AI-ready federal datasets; no new AI regulator |
| VI | Education and workforce | Build AI training into existing programs; study task-level job shifts |
| VII | Federal framework and preemption | Preempt state AI laws that impose undue burdens, with carve-outs |
The White House press release describes "six key objectives," matching sections I through VI, and then says the framework "can succeed only if it is applied uniformly across the United States," which is the subject of section VII.
What each section recommends
I. Children and parents. Congress should give parents tools to manage children's privacy settings, screen time and accounts; set "commercially reasonable, privacy protective" age-assurance requirements, such as parental attestation; require features that reduce risks of sexual exploitation and self-harm for minors; and affirm that existing child privacy protections apply to AI. It also asks Congress not to preempt state laws protecting children, such as bans on AI-generated child sexual abuse material.
II. Communities. In line with the Ratepayer Protection Pledge, Congress should ensure residential ratepayers do not pay more for electricity because of new AI data centers, while streamlining federal permitting so developers can build on-site power. It also calls for stronger action against AI-enabled scams targeting groups such as seniors, technical capacity in national security agencies "to understand frontier AI model capabilities," including through consultation with frontier developers, and AI resources for small businesses.
III. Intellectual property. The document says the administration believes training on copyrighted material does not violate copyright law, but acknowledges contrary arguments and supports letting the courts resolve it. It suggests Congress consider collective licensing without antitrust liability, without deciding whether licensing is required, and a federal protection against unauthorized AI digital replicas of a person's voice or likeness, with exceptions for parody, satire and news.
IV. Free speech. Congress should prevent the federal government from coercing technology providers, including AI providers, to ban, compel or alter content "based on partisan or ideological agendas," and give Americans a way to seek redress.
V. Innovation. Recommendations include regulatory sandboxes, AI-ready federal datasets for industry and academia, and a statement that Congress "should not create any new federal rulemaking body to regulate AI," relying instead on existing sector regulators and industry-led standards.
VI. Workforce. Congress should use non-regulatory methods to build AI training into education and apprenticeship programs, expand study of "task-level workforce realignment driven by AI," and strengthen AI programs at land-grant institutions.
VII. Preemption. Congress should preempt state AI laws "that impose undue burdens" to create "a minimally burdensome national standard." The framework says that standard should not preempt states' general laws protecting children, preventing fraud and protecting consumers, state zoning decisions about where AI infrastructure goes, or rules about a state's own use of AI. It says states should not regulate AI development, which it calls "an inherently interstate phenomenon," should not unduly burden uses of AI that would be lawful without AI, and should not penalize developers for a third party's unlawful use of their models.
What the framework does and does not do
- It does not change any law by itself. It is a set of recommendations to Congress. The White House said it looked forward to "working with Congress in the coming months to turn this framework into legislation."
- It does not create a new AI regulator. It explicitly recommends against one.
- It does not set licensing or safety-testing rules for frontier models. Its frontier-model recommendation is about government technical capacity and consultation. Separate federal actions, such as Executive Order 14409 of June 2026, deal with voluntary pre-release access for highly cyber-capable models; see our SI timeline.
On the official sources we checked, we did not find an enacted federal statute implementing the framework as of October 5, 2026.
How analysts have read it
Georgetown's Center for Security and Emerging Technology, a nonpartisan research center, noted that the framework goes further than earlier White House AI documents by explicitly calling on Congress to legislate, and that some proposals leave key details to Congress. It pointed out that Senator Marsha Blackburn released a legislative discussion draft, the TRUMP AI AMERICA Act, two days earlier, with some shared priorities and some differences, including on preemption. CSET suggested recommendations such as tracking task-level workforce changes and building national security agencies' capacity on frontier AI could draw bipartisan support.
Who this matters to
- AI developers and deployers: the preemption section would shape which state laws apply to products such as ChatGPT, Claude, Gemini and Grok.
- State policymakers: the carve-outs define what states would keep under the proposal.
- Parents and schools: section I is the most detailed on child safety.
- Policy watchers abroad: government AI testing bodies such as the UK AI Security Institute take a different, testing-led approach, which makes the US proposal's focus on preemption and sector regulators a useful contrast.
For the terminology change itself, see what SI means, our SI vs AI checklist for businesses and our superintelligence executive order guide. For the federal AI-for-science program the fact sheet lists alongside this framework, see Genesis Mission explained.
FAQ
What is the National AI Policy Framework?
It is a White House document released on March 20, 2026, titled "A National Policy Framework for Artificial Intelligence: Legislative Recommendations." It asks Congress to pass a national AI law covering child safety, communities, intellectual property, free speech, innovation, workforce and preemption of state laws.
Is the National Super Intelligence Policy Framework a different document?
We found no separate document. The September 29, 2026 fact sheet uses that name for a framework with legislative recommendations unveiled in March 2026, which matches the National Policy Framework for Artificial Intelligence.
Is the framework law?
No. It is a set of recommendations to Congress. It would need to be enacted as legislation to take effect.
Would it ban state AI laws?
It recommends preempting state AI laws that impose undue burdens, and says states should not regulate AI development. It would keep states' general consumer-protection, fraud and child-protection laws, zoning powers over AI infrastructure, and rules for states' own use of AI.
Does the framework create a new AI agency?
No. It says Congress should not create any new federal rulemaking body to regulate AI and should rely on existing regulators with subject-matter expertise and on industry-led standards.