Two federal blueprints for AI governance landed within 48 hours of each other this week. Neither is law. Both signal where the legislative fight is heading.
Status: These are proposals, not enacted legislation.
The White House released its National AI Legislative Framework on March 20, 2026, pursuant to President Trump’s Executive Order of December 11, 2025. At four pages, it is a principles document, not a bill. It urges Congress to build a unified federal regime, one that preempts the growing patchwork of state AI laws and keeps regulatory requirements minimal.
Sen. Marsha Blackburn (R-TN) moved two days earlier, releasing the TRUMP AMERICA AI Act as a nearly 300-page discussion draft on March 18. The name’s “4 Cs” framing, children, creators, conservatives, and communities, telegraphs its audience. It covers child safety, free speech, and intellectual property, and it proposes the same core objective as the White House framework: a national AI standard that displaces state-level regulation.
Where they agree and where they don’t:
| Dimension | White House Framework | Blackburn Bill | |—|—|—| | Federal preemption of state AI laws | Yes, urges Congress to adopt | Yes, proposes to establish | | New federal AI regulatory body | No, directs oversight to existing agencies | No (per available analyses) | | Copyright / AI training question | Defers to courts (ongoing judicial resolution) | Would specify AI training on copyrighted works is not fair use (per early legal analyses) | | Document length | 4 pages | ~300 pages | | Binding status | Legislative recommendation | Discussion draft |
The copyright divergence is the fault line. Sullivan & Cromwell’s analysis of the White House Framework confirms the administration’s position explicitly: the document leaves the question of whether AI training on copyrighted content constitutes infringement to ongoing judicial resolution rather than recommending a legislative answer.
Blackburn’s draft goes the other direction. According to early legal analyses, it would specify that unauthorized use of copyrighted works to train AI models does not qualify as fair use, a direct departure from the White House’s hands-off stance on the same question. That provision is a `[LEGAL-INTERPRETATION]` of the draft bill’s text, not a direct quote from it, and it remains subject to revision.
Why the preemption push matters now. Multiple states have enacted or are advancing AI-specific legislation, a fragmented landscape that both frameworks treat as a problem. Federal preemption would establish a single compliance baseline and override state rules in the covered areas. The Framework reportedly includes preemption carveouts in areas such as child safety and consumer protection, though the specific scope has not been independently verified against the document text.
What to watch: whether either document advances to formal legislative proceedings, and whether the copyright provision survives negotiation if the Blackburn bill reaches committee markup. The White House framework’s four-page outline will likely need to be translated into bill text before Congress can act on it.
TJS synthesis: The more significant development here is not that two frameworks appeared, it’s that they appeared together and largely agree on preemption while splitting on the question that will define AI development economics for the next decade. The copyright fault line is where the real legislative fight will happen. Compliance teams tracking federal AI legislation should monitor both tracks, but neither document creates obligations today. For guidance on building your compliance program before Congress decides, see our earlier analysis.