California's AI Transparency Act took effect into the teeth of a federal preemption push
The state's disclosure regime became operative on 2 August. The federal executive order signed in December directs a litigation task force at exactly this class of state law and conditions some grant funding on states declining to pass more of it.
Two governments now claim authority over the same disclosures. California requires them; the federal executive order "Ensuring a National Policy Framework for Artificial Intelligence" seeks a minimally burdensome national standard that preempts conflicting state rules, and stands up a task force to challenge them in court.
For a developer the practical position is unchanged: comply with California, because an executive order directing litigation is not a ruling, and the compliance deadline is now while the case is not. The same logic applied to the EU transparency obligations that came into force the same week.
The second-order effect is already visible in what states chose to pass this year. Preemption pressure appears to have pushed legislatures toward child safety, data centres and consumer protection — areas where federal appetite to litigate is lowest — and away from developer obligations. Colorado went further and repealed its own AI Act rather than defend it.
What the fight has not produced is a federal standard. Preempting state law requires legislation, not an order, and Congress has not passed any. The current equilibrium is a patchwork the federal government is suing to dismantle without having drafted a replacement.
Latham & Watkins — AI Executive Order Targets State Laws and Seeks Uniform Federal Standards → · Tech Policy Press — Where State AI Legislation Stands Half Way Into 2026 →