A modest AI transparency bill in Utah has become the latest flashpoint in an escalating conflict over who should regulate artificial intelligence in the United States: states, Congress, or the White House.
Earlier this year, Utah state Representative Doug Fiefia, a former Google executive now running for state Senate, introduced the Artificial Intelligence Transparency Act, a narrowly tailored proposal targeting “frontier” AI developers. The bill would have required companies building large-scale models to publish safety and child-protection plans and included whistleblower safeguards for employees raising concerns. By most legislative standards, it was limited in scope, applying only to systems trained at the highest computational thresholds and capping penalties at $1 million. It advanced unanimously out of committee.
Then it ran into the Trump administration.
In February, the White House Office of Intergovernmental Affairs sent a blunt letter to Utah lawmakers declaring the bill “unfixable” and incompatible with the administration’s AI agenda. Federal officials had already urged Fiefia, a Republican, to abandon the proposal in private conversations. With no path forward, the bill died in the state Senate.
The episode, highlighted in a recent report by The Next Web, illustrates a broader and intensifying conflict. States are moving aggressively to regulate AI, while the Trump administration is pursuing a coordinated strategy to block them in favor of a single, “minimally burdensome” national framework.
That federal strategy has unfolded in three stages. First, an executive order issued in December 2025 established an AI Litigation Task Force within the Department of Justice to challenge state laws in federal court. It also directed federal agencies to identify “burdensome” state regulations and tied certain federal funding streams to states’ willingness to avoid restrictive AI rules.
Second, the Commerce Department delivered a report highlighting state laws—particularly those in California, Colorado, and New York—for scrutiny, setting the stage for expected litigation beginning in 2026. Third, the administration released a sweeping national AI policy framework urging Congress to formally preempt state laws that create what it views as a fragmented and innovation-stifling regulatory landscape.
Administration allies have framed the issue in stark terms, per TNW. Former AI and crypto czar David Sacks warned that a patchwork of 50 state regimes would impose untenable compliance burdens on developers and raise constitutional concerns, particularly around free speech.
Read more: Utah Wants to Solve Washington’s AI Deadlock. Here’s How.
Statehouse legislative activity has nonetheless surged dramatically, from fewer than 200 AI-related bills introduced in 2023 to more than 1,200 in 2025, with 145 enacted into law. In just the first two months of 2026, dozens of chatbot safety bills were filed across more than half the states. Major economies including California, Texas, and Colorado have already enacted or are preparing to implement comprehensive AI statutes addressing transparency, governance, and algorithmic discrimination.
The momentum reflects a bipartisan consensus among state policymakers that federal inaction makes local action necessary. Utah Governor Spencer Cox, also a Republican, has argued that innovation and regulation are not mutually exclusive, particularly when it comes to protecting children and consumers from emerging AI risks.
That position has also found support among state attorneys general. A bipartisan coalition of 36 AGs has formally opposed federal preemption, warning that risks such as scams, deepfakes, and harmful AI interactions require tailored state-level responses.
Meanwhile, Congress remains gridlocked. Efforts to impose a federal moratorium on state AI laws were decisively rejected, most notably in a 99–1 Senate vote stripping such a provision from major legislation. Even as lawmakers debate comprehensive frameworks—including a sweeping draft bill that would impose duties of care and create new liability regimes—no consensus has emerged.
The result is a three-way standoff: states are legislating, the executive branch is preparing to litigate, and Congress is struggling to arbitrate.
In the absence of federal legislation, AI governance in the United States is increasingly being shaped not by statute but by executive action, court challenges, and political pressure from both industry and public-interest groups. Tech companies and aligned political action committees have poured hundreds of millions of dollars into influencing the outcome, while pro-regulation advocates are mobilizing their own resources in response.
For lawmakers like Fiefia, the stakes are both immediate and structural. His failed Utah bill has become a symbol of a larger question: whether states will retain the authority to regulate a transformative technology in the absence of federal rules—or whether that authority will be preempted before it can fully take shape.
