A policy of no rules turns out to produce its own kind of rules. They are just harder to read, harder to cite, and impossible to challenge through normal administrative channels. That is the practical situation facing anyone building or deploying AI systems in the United States right now, and it is worth understanding precisely because the official story is almost the exact opposite of what is actually happening. ## The Stated Position Versus the Operational Reality The Trump administration entered office opposing AI regulation, according to reporting by Axios on the administration's approach to AI governance. The premise was that lighter federal oversight would accelerate innovation and preserve American competitiveness. What Axios describes instead is a White House that is actively shaping the AI industry through case-by-case interventions, without codifying those interventions as clear rules. That distinction matters enormously for anyone trying to build a compliant product. A formal rule is published, debated, and legally challengeable. An informal intervention is none of those things. It exists as precedent only if you were in the room, and as uncertainty for everyone else. This is not a trivial semantic difference. When governance operates through ad hoc decisions rather than durable written standards, builders face what compliance professionals sometimes call a navigational vacuum: they know the government has opinions, but they cannot read those opinions in advance, cannot rely on them as stable guidance, and cannot point a counterparty to a regulation that explains the constraint. The result, as Axios frames it, is major uncertainty, the kind that chills product decisions not because anything is technically prohibited but because no one is confident about what the next intervention will target. ## The Preemption Move and What It Left Behind The structural backdrop to this situation is the December 11, 2025 Executive Order titled "Ensuring a National Policy Framework for Artificial Intelligence," signed by President Trump and published by the White House. As analyzed by Epstein Becker Green, the EO aims to prevent a patchwork of state laws by restricting states from independently regulating AI in ways the administration characterizes as "onerous and excessive" or conflicting with federal priorities including innovation and global competitiveness. The stated goal is to reduce what the EO calls "cumbersome" state regulation that could "stymie innovation." The Heinrich Böll Stiftung analysis of the same period identifies the resulting gap directly: the Trump administration's EO seeks to preempt state laws without a federal framework already in place, leaving a regulatory void. That void is not empty, however. It is filled by the informal interventions Axios documents. So the practical compliance environment for an AI builder in the United States currently looks like this: state-level protections are under pressure from federal preemption, a replacement federal framework has not been enacted, and White House influence is being exercised deal by deal and case by case. Builders are left navigating through inference rather than instruction. ## What the Global Picture Adds The United States is not operating in isolation, and that cross-border dimension is where the compliance pressure compounds for builders with any international reach. The Heinrich Böll Stiftung notes that both the U.S. and the EU have moved toward deregulation in an attempt to attract investment and foster innovation, with the EU also scaling back the AI Act. But the EU's rollback is happening within an existing, published legal structure, with known timelines for which provisions apply to which systems. The U.S. shadow policy has no equivalent transparency. For builders serving learners in multiple jurisdictions, that asymmetry creates a concrete problem. The EU AI Act Newsletter, published by the Future of Life Institute, notes that Trump's AI Action Plan prioritizing American supremacy over risks will not protect U.S. companies from global regulation. A U.S.-based edtech platform deploying an AI tutoring or assessment system for European users still faces EU obligations, regardless of what informal signals come out of Washington. The practical implication is that compliance teams cannot treat the absence of a formal U.S. rule as a green light. They have to model the informal U.S. environment alongside the formal international one, simultaneously. ## What Builders and Learners Should Actually Do With This Understanding informal governance environments is a genuine skill, and it is one worth developing now. The first step is recognizing that "no regulation" and "no regulatory risk" are different claims. The Trump administration's posture, as documented by Axios, is producing de facto constraints through intervention rather than through statute. That means the compliance question is not only "what does the law require" but also "what has the administration signaled it will act on, and how do I document my reasoning if I am asked to justify a product decision." For those learning AI governance, the current U.S. environment is actually a useful case study in how policy can operate outside the traditional notice-and-comment rulemaking process. Brookings has tracked the AI policy directions of the Trump administration as a matter of formal analysis, and the picture that emerges is one where the architecture of governance is being constructed informally, in real time. That is not a comfortable situation for builders who need stable ground, but it is comprehensible once you understand the mechanism. Watch for whether the June 3, 2026 Executive Order on AI innovation and security, signed by President Trump as reported by Portland State University's technology law resource, produces implementing guidance that firms up any of these informal signals into written obligations. If it does, the navigational map improves. If it does not, the pattern Axios identified will continue to be the operating environment, and reading it carefully will remain the most useful compliance skill available. ## Sources - Trump's shadow AI policy

Sources