The US is moving toward a more centralised approach to AI regulation. An Executive Order issued in December 2025 directs the Attorney General to identify and challenge state laws that are inconsistent with the Administration’s national AI policy framework. The order accelerates a preemption debate that has been building since states began passing their own AI bills.
For UK and international firms operating in the US, the practical question is whether the regulatory landscape will become simpler or more contested in the next few years.
The preemption problem
Over the past two years, several US states have enacted AI-specific legislation. Colorado’s AI law, California’s transparency and safety bills, and various state-level deepfake and biometric rules have created a patchwork of obligations. For businesses operating nationally, this patchwork increases compliance cost and legal uncertainty.
The Executive Order reflects a view that this patchwork obstructs national AI leadership. By directing the Department of Justice to challenge conflicting state laws, the Administration is signalling that AI policy should be set at the federal level. The order does not automatically invalidate state laws, but it makes litigation and federal preemption arguments more likely.
What it means for operators today
The immediate effect is limited. Executive Orders do not override statutes, and state laws remain enforceable until a court rules otherwise or Congress passes federal legislation. However, the order changes the enforcement environment. State attorneys general and legislatures may become more cautious about expanding AI rules if they expect federal challenges.
For operators, the main action is to keep tracking both federal and state developments. A federal preemption strategy could eventually reduce the number of regimes you need to comply with. In the short term, it may create more litigation and uncertainty as courts work out which laws survive.
Implications for UK firms
UK firms selling AI services into the US should not assume that federal preemption will produce a single, simple rulebook. The US system leaves significant room for state consumer protection, privacy and sectoral laws to continue applying. Preemption arguments will turn on the specific language of each law and the constitutional basis for federal authority.
The more useful reading is strategic. The Executive Order is part of a broader US effort to position itself as the leading jurisdiction for AI development by reducing regulatory friction. That may influence where AI companies choose to incorporate, where they locate research and how they design compliance programmes.
The bottom line
Federal preemption of state AI laws is now an explicit US policy objective, not just a legal theory. For UK operators, the near-term task is unchanged: comply with the state laws currently in force while watching federal action closely. The longer-term bet is whether a national framework emerges that simplifies US compliance, or whether the next few years bring more conflict between federal and state regulators.