Three isn’t a coincidence.
In December 2025, the Trump Administration issued an AI Preemption Executive Order, a signal of direction, not a compliance framework. In the months that followed, Senator Blackburn introduced the TRUMP AMERICA AI Act in Congress. On March 20, 2026, the White House’s own Office of Science and Technology Policy released its National Policy Framework for AI. Three federal AI governance actions. Three different mechanisms. One underlying goal: replace the fragmented state-by-state AI regulatory landscape with federal authority.
What compliance teams don’t yet have is clarity on which track succeeds, or whether any of them do.
The Three Frameworks, Precisely
Understanding the compliance stakes requires treating these three tracks as distinct instruments, not interchangeable versions of the same initiative.
Track 1: The OSTP Framework (Executive Branch, March 20, 2026)
The White House OSTP framework is a legislative wish list from the executive branch. Per analysis by Covington & Burling, it contains more than two dozen recommendations to Congress, proposals, not mandates. The framework recommends consolidating AI oversight within existing sector-specific agencies (FDA, FTC, SEC, and analogues), explicitly rejecting the creation of a new AI regulatory body. According to legal analysts who reviewed the document, its recommendations reportedly address child safety, free speech, intellectual property, and workforce protection.
The OSTP framework also signals alignment with President Trump’s December 2025 AI Preemption Executive Order and July 2025 AI Action Plan, confirming that preemption is an administration-wide priority, not a single-agency position.
Track 2: The Blackburn TRUMP AMERICA AI Act (Congressional, ongoing)
The Blackburn bill is actual legislation moving through Congress. This hub’s prior coverage details its structure, the short version is that it names specific state AI laws and seeks to preempt them through direct federal legislation. Unlike the OSTP framework, the Blackburn bill is a congressional instrument. It requires a legislative majority, not executive implementation.
The Covington analysis covered both the OSTP framework and the Blackburn bill in a single piece, a signal that policy analysts see them as complementary rather than competing. Whether Congress agrees is a separate question.
Track 3: The Status Quo (Sector Agencies, Ongoing)
The third track gets the least attention because it doesn’t produce press releases. Existing sector-specific federal agencies, the FTC, FDA, CFPB, and others, already have enforcement authority over AI applications in their regulated domains. They’re using it. The FTC has investigated algorithmic discrimination. The FDA is developing AI medical device guidance. The CFPB has issued guidance on AI-driven credit decisions.
This track is slower and more fragmented than either preemption vehicle. It’s also the only one that’s currently operational.
The Preemption Question
Both Track 1 and Track 2 share the same goal: preempt state AI laws. The practical stakes of that goal are significant. At least 20 US states have passed or advanced AI-specific legislation as of early 2026. California’s SB 1047 debate, Colorado’s AI Act, Texas’s AI regulatory proposals, and a growing cluster of state privacy laws with AI provisions all represent compliance obligations that federal preemption could eliminate, or replace with something different.
The preemption framing is not “federal regulation vs. no regulation.” It’s “federal uniformity vs. state patchwork.” For organizations operating nationally, federal uniformity sounds appealing. But the OSTP framework and Blackburn bill don’t promise uniformity that matches what states currently require. They promise federal uniformity on federal terms, which may be more permissive in some areas and more restrictive in others.
The “No New Agency” Consensus and What It Means for Enforcement
Both the OSTP framework and the Blackburn bill reject creating a new dedicated AI regulatory body. That consensus is more significant than it might appear. A purpose-built AI regulator would have AI expertise as its core competency and AI compliance as its only mandate. Routing enforcement through existing agencies means AI compliance questions go to bodies whose primary missions, staffing, and enforcement cultures were built for different problems.
Per Freshfields LLP’s analysis of the OSTP framework, this is a deliberate choice, not an oversight. The rationale is that sector-specific expertise in existing agencies is more valuable than new AI-specific expertise at a new agency. Whether that rationale holds under enforcement pressure is one of the open questions the next two years will answer.
For compliance teams, the practical implication is straightforward: if federal preemption passes in either form, your compliance touchpoints will be industry-specific. A healthcare AI vendor answers to FDA. A lending platform’s AI decisions face CFPB scrutiny. An advertising technology company’s AI practices are FTC territory. The vertical matters more than the “AI” category.
Stakeholder Positions
The stakeholder map on federal AI preemption is not simple, and this brief will not fabricate positions where they aren’t verified. What the Filter’s sources confirm:
Major AI companies broadly favor federal regulatory uniformity over state patchwork, the compliance cost of 50 different state frameworks is a genuine operational burden. This is a documented industry position, not an inference.
State attorneys general and state legislators in states with active AI legislation have documented opposition to federal preemption, preserving state authority over consumer protection and civil rights applications of AI is the consistent stated rationale.
The civil liberties community is divided. Federal preemption may eliminate state-level protections that are stronger than what the OSTP framework or Blackburn bill would replace them with. That’s a documented concern in the policy literature, though specific organizational positions on the March 2026 documents are not confirmed in this cycle’s source material.
Congress is not a unified actor on this question. The Blackburn bill has co-sponsors but faces opposition from legislators who prefer either stronger federal AI regulation or the preservation of state authority. The OSTP framework is a presidential recommendation to a Congress that doesn’t have to follow it.
What Compliance Teams Should Do Now
Waiting for clarity is a compliance strategy, it’s just not a good one. Here’s what the current landscape actually supports:
Map your state AI exposure. If federal preemption passes, what state obligations change? If it doesn’t pass, what’s your current compliance gap? You need this map either way. Build it now.
Watch the OSTP primary document. The White House framework’s actual text, not law firm summaries of it, is the authoritative source for what the administration is recommending. If it becomes publicly accessible, read it directly. Law firm analysis is useful interpretation; it’s not a substitute for primary source review.
Don’t plan for one outcome. Three tracks are live simultaneously. Compliance programs that assume “federal preemption will pass as proposed” are making a bet. A scenario-based approach, what do we do if Track 1 passes, if Track 2 passes, if neither passes, is more defensible.
Treat sector-agency enforcement as current risk. Track 3 isn’t waiting. FTC, FDA, and CFPB enforcement is happening now, in your sector, under existing authority. Federal preemption doesn’t change that picture in the near term.
TJS Synthesis
The OSTP framework, the Blackburn bill, and the existing sector agency landscape aren’t three versions of the same story. They’re three different answers to the same question: who has authority over AI in the United States, and on what terms?
The administration’s answer, centralize through existing agencies, preempt state law, no new body, is now visible in both the executive branch and the legislative track it’s sponsoring. Whether Congress delivers that answer is genuinely uncertain. What isn’t uncertain: compliance programs that treat the federal preemption question as settled in either direction are running ahead of the evidence.
The organizations that will handle this best are the ones building compliance architectures flexible enough to pivot when one of these tracks actually lands.