In late 2025, the struggle over how to regulate artificial intelligence in the US entered a new and volatile phase. After Congress twice failed to pass a law banning state artificial intelligence laws, President Donald Trump on December 11 signed a broad executive order designed to limit states from regulating the industry and to promote a “minimally burdensome” national policy. The order, a partial win for major technology companies that oppose a patchwork of state rules, sets up a direct clash between the White House and state governments at a moment when public concern about artificial intelligence’s impact on jobs, mental health, and the environment is rapidly growing.
The executive order instructs the Department of Justice to form a task force to sue states whose artificial intelligence laws conflict with Trump’s light-touch vision, and tells the Department of Commerce to withhold federal broadband funding from states if their artificial intelligence regulations are considered “onerous.” Legal experts expect the order to be used against a limited set of provisions, particularly those addressing transparency and bias, which are more common in Democratic states. Yet states such as New York and California have already enacted significant measures, including New York’s Responsible AI Safety and Education Act and California’s frontier artificial intelligence safety law SB 53, which require companies to publish safety protocols, report critical incidents, and guard against catastrophic harms like biological weapons or cyberattacks. These laws, tempered by heavy industry lobbying, could become central test cases if the administration moves to challenge them.
As 2026 approaches, the conflict is shifting into the courts and the political arena. Democratic-led states, and some Republican states with vocal artificial intelligence regulators, are expected to resist federal attempts at preemption, even as other Republican states may retreat for fear of losing broadband funds or provoking the administration, potentially chilling new lawmaking. In Congress, efforts to craft federal artificial intelligence legislation have stalled after failed attempts to attach a moratorium on state laws to tax and defense bills, and Trump’s executive order appears to have hardened partisan divisions rather than foster compromise. Outside Washington, litigation and state legislation will proliferate, particularly around child safety, with lawsuits against chatbot providers, a California ballot initiative backed by OpenAI and Common Sense Media, and proposed rules on data center power and water use. At the same time, rival super PACs funded by technology leaders and artificial intelligence regulation advocates are preparing to pour tens of millions into 2026 races, underscoring how decisions in state capitals may shape the trajectory of artificial intelligence far beyond US borders for years to come.
