Utah built a shield against AI harm; the White House knocked it down

SALT LAKE CITY — For the better part of this legislative session, Utah looked like it might become the most important state in America when it comes to regulating artificial intelligence.

It had the bills. It had the momentum. It had legislators willing to take on some of the world's most powerful technology companies. And then, in the final days of the session, it hit a wall — one built in Washington, D.C.

HB286, the centerpiece of Utah's AI accountability push this session, has been stalled after the White House came out strongly against it, calling the bill "unfixable." The legislative session closed on March 6. For now, the bill is dead.

For Utahns who were counting on new rules to govern how AI systems affect their lives — their jobs, their health care, their children's schools — that's a significant setback. But it's not the end of the story.

What HB286 would have done

To understand why the bill's stall matters, it helps to know what it actually contained.

HB286 targeted what legislators called "frontier models" — the most powerful AI systems, the kind built by companies like OpenAI, Google and Anthropic. Under the bill, companies deploying those systems in Utah would have faced mandatory transparency requirements and fines of up to $3 million for violations that caused harm or operated outside defined ethical guardrails.

The bill was part of a broader dual-track approach Utah has been developing: strict accountability for the biggest AI players, paired with a more supportive framework for smaller, local businesses trying to innovate responsibly.

That second track — the friendlier one — is the Learning Lab, a state-sanctioned sandbox where Utah companies can test high-stakes AI applications under state oversight without immediately facing the fines and liabilities that might otherwise discourage experimentation. Think of it as a "try before you're regulated" lane, specifically designed so that a small Utah startup doesn't face the same compliance burden as a trillion-dollar tech giant.

Utah's Companion Chatbot Safety Act, which passed with a 68-1 vote, did survive the session. That bill focuses on protecting users — particularly vulnerable ones — from AI companionship apps that could cause psychological harm. It's a narrower protection, but a real one.

Why the White House pushed back

Federal opposition to HB286 reflects a broader tension playing out across the country right now.

The current administration has signaled it wants AI regulation handled at the federal level — not state-by-state. The argument is that a patchwork of 50 different state AI laws would create compliance chaos for technology companies and slow down American AI development at a moment when the U.S. is competing directly with China.

Critics point out that "we'll handle it federally" can mean "we won't handle it at all" — at least not quickly. Congress has been discussing federal AI regulation for years, yet no comprehensive legislation has been passed.

In the meantime, the AI systems the bill was designed to govern are already operating in Utah. They're already making decisions that affect Utahns. The gap between the technology's reach and the law's reach keeps widening.

Rep. Doug Fiefia, R-Herriman, said HB286 was "meant to start an important conversation about AI transparency and how Utah should approach this rapidly evolving technology."

"We'll continue working with industry and stakeholders during the interim to refine the approach so Utah can support innovation while also putting thoughtful protections in place for consumers," he continued.

What this means for Utah businesses

For the Utah companies that were tracking HB286 — particularly those building or deploying AI in regulated industries like health care, legal services and finance — the bill's stall creates a specific kind of uncertainty.

Shawn Miele, CEO at MyAdvice, a Park City-based company that builds AI systems for doctors, dentists and lawyers across the country, has been watching Utah's regulatory approach closely. His take on the trust problem that regulation was meant to solve is direct: "There are segments of our potential clients that absolutely do not trust AI yet. And that trust issue isn't easy to solve. It's something you have to work on over time."

Regulation, done well, accelerates that trust-building. It gives businesses and consumers a shared framework for what "responsible AI" actually means — not just a company's marketing claim, but a legal standard. Without it, the burden falls entirely on individual companies to self-regulate, and on individual consumers to figure out who to believe.

The Learning Lab framework, which survived the session intact, does offer some of that structure for smaller Utah businesses. Companies that enroll can test their AI applications with state oversight and secure what are called Regulatory Safe Harbors — essentially a documented record that they operated in good faith under state guidance. That's meaningful protection, even without the broader frontier model accountability that HB286 would have established.

What comes next

The legislative session is over, but the conversation isn't. Utah's legislators have been among the most active in the country on AI policy — this is the same state that passed the country's first AI disclosure law in 2023 — and the people who worked on HB286 are not going away.

The more likely outcome is a revised bill next session, potentially narrowed in scope to address the White House's specific objections while preserving the core accountability framework. Utah has navigated federal-state tensions on tech policy before, and the state's track record suggests it will find a path.

In the meantime, the $100 million AI infrastructure initiative anchored here — built around NVIDIA computing resources and designed to give Utah's small and mid-sized businesses access to enterprise-grade AI capabilities — continues moving forward. The pro-human philosophy that has defined Utah's approach to AI this session isn't gone. It just didn't make it into law this time.

For Utahns paying attention, that distinction matters. The question of who gets to set the rules for AI — states, the federal government or the technology companies themselves — is one of the defining policy fights of the next decade. Utah just had a front-row seat to Round 1.

Read more

Artificial Intelligence in Utah

For more news about Artificial Intelligence and its development and advancement in society, visit this section on KSL.com.

Related topics