Republicans in the Utah Legislature are advancing the “AI Transparency Act,” a measured bill that tries to impose basic accountability on the biggest, most powerful AI developers in the world.
The bill, H.B. 286, does not create a new regulatory agency. It does not micromanage technical design choices. It does not hand trial lawyers a private right of action. It focuses on transparency, safety planning, reporting serious safety incidents, and protecting whistleblowers who sound the alarm when something goes wrong.
And yet the White House has tried to crush it. The administration sent Utah leaders a terse message declaring blanket opposition to the bill, calling it “unfixable” and claiming it conflicts with the administration’s AI agenda.
That posture tracks with the approach being driven by the administration’s top AI and crypto official, David Sacks, who has pushed aggressively against state-led AI policymaking. Whether the pressure campaign is explicit or simply implicit, the effect is the same: Republican states are being told to stand down and let Washington handle it.
That is the wrong approach, and conservatives should say so.
The White House is right to worry about clumsy, heavy-handed government interference in artificial intelligence. AI could turbocharge productivity, open new frontiers in medicine and science, and usher in a new economic golden era for the United States. Conservatives should be skeptical of the kinds of AI rules coming out of Europe and deep-blue states, where “safety” often becomes a pretext for speech policing, ideological enforcement, and permanent bureaucratic control.
But there is a difference between resisting left-wing AI regulation and declaring war on any meaningful guardrails at all.
The fastest way to misunderstand this fight is to treat Utah’s bill like it is a progressive or socialistic attempt to micromanage algorithms or enforce political dogma. Utah is not trying to turn AI into a speech code. It is not trying to impose a European-style compliance regime. It is trying to force the largest AI companies to level with the public about serious risks, to think through safety in advance, and to report major safety incidents when they happen.
In other words, it is an attempt to do something conservatives usually claim to support: require powerful institutions to be transparent and accountable when their products can cause real harm.
If the bill is truly flawed, the responsible move is to propose fixes, such as clarifying definitions or adjusting penalties. But waving it away as “unfixable” is not good policymaking.
The deeper problem is that the Trump administration’s rejection of state authority when it comes to AI is being sold as pro-liberty, even though it could easily produce the opposite result. Sometimes, making regulatory changes is necessary to protect freedom, especially when markets are already heavily regulated. This is especially true when it comes to AI companies, which must comply with numerous rules and regulations imposed by foreign governments, many of which do not support the freedoms protected in the U.S. Constitution.
For example, the European Union and China have built strict AI regulatory regimes in recent years, and multinational corporations will often comply with the most restrictive rules, because they operate across borders and want one set of standards. When those standards are written by people who do not share American traditions of free speech, due process, and limited government, Americans could end up living under de facto policies created outside their constitutional system.
Even if governments were to stay on the sidelines, corporate compliance departments would then become the chief regulators, and many of them do not value free speech and other liberties, either. And don’t be fooled into thinking they operate within a free market. Large tech corporations benefit from limited liability and countless other legal protections that make it difficult for regular people to hold them accountable.
The point is, when it comes to emerging technologies such as AI, somebody will decide what is allowed, what is restricted, what is flagged, what is promoted, and what is buried. Conservatives of all people should understand this. We just lived through an era in which private corporations shaped public life, throttled narratives, and punished dissent on countless social media platforms.
AI models will magnify that dynamic. These systems can curate information, summarize news, recommend actions, filter content, and act as gatekeepers for knowledge on a scale humanity has never before known.
Further, contrary to what Sacks and other opponents of regulating AI claim, states are not an obstacle to liberty. They are a primary safeguard of it. The constitutional case is simple. The 10th Amendment reserves to the states, or to the people, the powers not delegated to the federal government. Consumer protection, child safety, fraud prevention, and civil enforcement have long been part of the states’ traditional powers. States regulate products and services within their borders every day because they have an obligation to protect residents from foreseeable harms.
There is no reason to think the federal government will do a better job of regulating AI, or even that it will soon regulate AI meaningfully at all. Despite the AI industry’s tremendous growth in recent years, there are still no substantial federal AI safeguards in place.
TRUMP AND DESANTIS LOCK HORNS IN GOP SPLIT OVER AI
Conservatives and the White House should not confuse “limited government” with “no rules for the biggest players.” And they should not pretend concentrated private power leads to more freedom, especially when the interests involved are already benefiting from government protections and often work in close coordination with lawmakers and government bureaucracies, both in America and abroad.
Utah lawmakers seeking to enact common-sense reforms should be celebrated, not chastised. They are trying to lead with a restrained, transparency-first approach. The White House should let them.
Justin Haskins is a New York Times bestselling author, vice president at the Heartland Institute, and a senior fellow for Our Republic.