White House Sends Draft Federal AI Framework to Congress, Seeks Preemption of Most State AI Laws
The White House has shared a draft federal AI framework with congressional leadership, urging Congress to enact a single national "One Rulebook" this year that would preempt many state AI laws. The proposal says AI development is inherently interstate and tied to national security — limiting states' ability to regulate development or penalize developers for third-party misuse — while explicitly preserving traditional state police powers (child protection, anti-fraud and consumer-protection laws) and zoning authority over AI infrastructure.
📌 Key Facts
- Fox News reported it obtained the White House's draft legislative AI framework and said the document would be shared with congressional leadership on Friday.
- OSTP Director Michael Kratsios and White House AI and crypto official David Sacks gave on‑the‑record interviews saying the framework aims to create 'one national policy' or a single 'One Rulebook' for AI and would explicitly preempt many state AI laws.
- The draft framework argues AI development is 'inherently interstate' and linked to foreign policy and national security, and says states should not be allowed to regulate development or penalize AI developers for third parties’ unlawful uses of their models.
- The proposal specifies that federal preemption would not override states' traditional 'police powers' — including child‑protection, anti‑fraud, and consumer‑protection laws of general applicability — nor state zoning authority over AI infrastructure placement.
- The White House is urging Congress to codify the framework 'this year,' framing the proposal as capable of winning bipartisan support and as designed to prevent censorship while protecting free speech and children.
📊 Relevant Data
As of 2026, state AI laws in the US commonly address themes such as transparency, bias prevention, data privacy, and accountability, with examples including prohibitions on discriminatory AI in employment effective January 1, 2026.
Artificial Intelligence Regulations: State and Federal AI Laws 2026 — Drata
Facial recognition technology exhibits failure rates as high as 35% for Black faces, significantly higher than for other racial groups.
When the algorithm is wrong: A new partnership calls out racism in AI systems — University of Toronto
AI-driven hiring systems disproportionately downgrade resumes from Black candidates, contributing to racial disparities in employment opportunities.
Bias in AI-driven HRM systems: Investigating discrimination risks and mitigation strategies — ScienceDirect
Data centers are projected to consume 6.7% to 12% of US electricity by 2028, up from 4.4% in 2023, potentially increasing electricity prices.
Here's how AI data centers affect the electrical grid — CNN Business
Black and Latino households pay 13–18% more on average for energy per square foot of housing compared to White households.
China accounts for roughly 70% of global rare earth production and dominates processing, giving it leverage in AI technology supply chains.
Beijing's dominance in rare earth processing leaves others scrambling — Fortune
The US leads China in AI chip performance, with top US AI chips being five times more powerful than China's.
The Complicated Stakes of the AI Race Between the U.S. and China — Time
📰 Source Timeline (2)
Follow how coverage of this story developed over time
- Fox News Digital obtained the actual legislative framework document, not just descriptions from sources, and reports that it will be shared with congressional leadership on Friday.
- White House OSTP Director Michael Kratsios and AI & Crypto Czar David Sacks give on-the-record interviews explaining that the framework is meant to create 'one national policy' and a single 'One Rulebook' for AI, explicitly preempting many state AI laws.
- The framework states that states should not be allowed to regulate AI development because it is 'inherently interstate' and tied to foreign policy and national security, and that states should not penalize AI developers for third parties’ unlawful uses of their models.
- The proposal specifies that federal preemption should not cover states’ traditional 'police powers' such as child-protection, anti-fraud and consumer-protection laws of general applicability, nor state zoning authority over AI infrastructure placement.
- The article emphasizes that the White House wants Congress to codify the framework 'this year' and argues it can garner bipartisan support, framing it as designed to prevent censorship and protect free speech and children.