A quick note before the substance: this brief draws on legal analysis rather than the primary government document. The Trump Administration published “A National Policy Framework for Artificial Intelligence: Legislative Recommendations” on March 20, 2026. Mayer Brown law firm reviewed and summarized the framework. That analysis is the source for the policy-area descriptions below. Where the language says “the framework addresses” or “the framework would,” read it as “according to Mayer Brown’s review.” The primary government document URL is flagged for the resolve-urls stage.
These are legislative recommendations. Congress has not acted on these proposals as of March 25, 2026. Nothing described below is currently in effect as federal law.
With that established: the White House sent Congress a seven-area AI policy blueprint last week. According to Mayer Brown’s analysis, the framework covers: fostering AI innovation, promoting U.S. AI dominance globally, preempting certain state AI laws, preventing government coercion of AI providers, protecting free speech in AI contexts, pursuing a non-regulatory approach to AI workforce training, and carving out exceptions that preserve state authority over traditional police powers, children’s protection, fraud, and consumer protection.
That last point is worth unpacking. The framework would preempt state AI laws broadly – but Mayer Brown’s review indicates it explicitly does not extend preemption to states’ traditional regulatory domains. California’s consumer protection authority, for instance, appears preserved under this structure. What gets preempted is the patchwork of state-level AI-specific liability frameworks that companies have been navigating.
For compliance professionals, the practical read is this: nothing changes today. The framework is a Congressional ask, not a mandate. But the direction it signals is clear. The Administration is positioning itself against state-level AI regulation as a governing regime while simultaneously avoiding direct federal content moderation. That posture, if enacted, would centralize AI governance at the federal level while keeping the federal government out of algorithmic decisions.
This document follows earlier White House AI policy communications, including the Administration’s contested position on copyright and AI. That prior brief addressed copyright specifically; this framework covers a broader set of policy dimensions and represents a new, distinct document rather than a revision of earlier positions.
What to watch: whether any Congressional sponsor picks up these recommendations as legislation, and which provisions, preemption in particular, attract the most pushback. State attorneys general have been active in AI enforcement; preemption authority is the most legally contested element of the framework. Also watch whether the workforce provisions (non-regulatory, voluntary) draw a response from labor groups or from executives like Dimon who are calling for something more interventionist.
The federal AI regulatory picture is being built in parallel tracks: an innovation framework from the White House, a workforce literacy program from DOL, roundtables at Treasury, and Congressional inaction so far. That’s not a coherent strategy yet. It may become one. Or it may remain a set of disconnected signals that compliance teams have to map independently.