Trump's AI Framework Would Kill State AI Laws — Here's What It Actually Says
The White House released a national AI legislative framework on March 20 that would federally preempt all state AI regulation, shield developers from third-party misuse liability, and fast-track data center permitting. The Senate already killed one preemption attempt 99-1.
What Happened
On March 20, 2026, the Trump Administration released a national AI legislative framework calling on Congress to pass federal AI legislation that would preempt state AI laws, establish developer liability shields, fast-track data center permitting, and create online safety provisions for children.
The framework builds on Executive Order 14179 signed in December 2025, which established an AI Litigation Task Force within the Department of Justice and directed federal agencies to evaluate state AI laws. The new framework represents the administration's formal ask to Congress: convert these executive actions into permanent legislation.
The Six Pillars
The framework addresses six key areas:
1. Federal Preemption of State AI Laws — The centerpiece provision would bar states from regulating AI development, prevent states from imposing burdens on AI use for lawful activities, and prevent states from holding developers liable for how third parties misuse their models. This directly targets laws in Colorado, California, Texas, and over a dozen other states that have enacted AI-specific regulation.
2. Developer Liability Shields — AI model creators would be shielded from lawsuits related to downstream misuse of their models by third parties. This is the provision Big Tech has lobbied hardest for — it effectively means OpenAI, Anthropic, Google, and Meta cannot be sued when someone uses their models to cause harm.
3. Data Center Permitting Fast-Track — The framework calls for streamlined permitting requirements so data centers can generate power on site, reducing the multi-year approval processes that currently delay AI infrastructure deployment. As we covered in our analysis of AI's power crisis, the electrical grid needs 175 GW of additional capacity by 2033 just for data centers.
4. Child Safety Online — Requires AI platforms likely to be accessed by minors to implement features reducing potential sexual exploitation of children and discouraging self-harm. Calls for parental account controls and privacy protections.
5. Anti-Censorship Provisions — Prevents AI platforms from engaging in viewpoint-based content moderation, framed as protecting free speech but criticized by safety researchers as potentially weakening content moderation systems.
6. National Security — Maintains export controls on advanced AI chips and models to adversarial nations, continuing the Biden-era approach to AI technology transfer restrictions.
Why It Matters
The framework represents the most aggressive federal attempt to consolidate AI governance power. Three specific mechanisms are already in play:
- DOJ AI Litigation Task Force (operational since January 10, 2026) is actively evaluating state AI laws for potential federal court challenges
- Commerce Department was directed to evaluate state AI laws by March 11, 2026, with states deemed to have "onerous" laws losing eligibility for $42 billion in BEAD broadband funding
- FTC was directed to classify state-mandated bias mitigation as a "per se deceptive trade practice"
The Colorado Problem
The framework specifically cites the Colorado AI Act as requiring entities to "embed ideological bias within models." Colorado delayed enforcement of its AI Act from February 1 to June 30, 2026, as lawmakers consider revisions under federal pressure.
Colorado's law requires deployers of high-risk AI systems to conduct impact assessments and provide consumer notification when AI makes consequential decisions. The administration frames this as ideological overreach; Colorado legislators argue it is basic consumer protection.
Congressional Reality Check
The framework faces steep congressional headwinds. During the "One Big Beautiful Bill Act" debate, a provision that would have preempted state AI regulations for 10 years was stripped by a 99-1 Senate vote. Bipartisan opposition to federal preemption of state consumer protection laws runs deep, and the framework asks Congress to do exactly what it just refused to do.
Industry Reaction
Major AI companies have publicly supported the framework's liability shield and preemption provisions. OpenAI, which recently raised $110 billion at a $730 billion valuation, has lobbied extensively for federal preemption. Critics argue this amounts to the largest companies writing their own regulatory environment.
Civil society organizations and state attorneys general have pushed back, arguing that federal preemption would create a regulatory vacuum — eliminating state protections before any federal replacements exist.
What Comes Next
The administration wants Congress to convert the framework into a bill "in the coming months." Given the 99-1 preemption vote, this timeline is optimistic. More likely outcomes include:
- Piecemeal legislation addressing child safety (bipartisan support exists)
- Continued DOJ challenges to individual state laws in federal court
- Executive agency actions (FTC, Commerce) that don't require congressional approval
- States continuing to enforce existing laws while legal challenges proceed
The framework sets the terms of debate for what will likely be the defining AI governance fight of 2026-2027.
Sources: White House, CNBC, Ropes & Gray analysis