Homelessness has become a perennial issue in Utah. An annual statewide report released this August found that people experiencing homelessness increased by 18% over the previous year.
Those rising statistics have led local elected officials to put pressure on the state for more resources, as well as the announcement of a 1,300-bed homeless campus in Salt Lake City.
Beyond Utah, other places are harnessing new technology to tackle the issue.
Since 2022, Los Angeles County has been using a predictive AI model to help identify individuals at risk of falling into homelessness.
“Our question was, could we use data to better understand not just who is pressured economically, but who are the subset of households or people who are going to end up on the street or in a shelter?” said UCLA California Policy Lab Executive Director Janey Rountree. “If you want to prevent that specific outcome, you really need to have smart strategies for identifying who's at most risk of ending up on the street or in a shelter.”
In October, the UCLA California Policy Lab and Los Angeles County were recognized as finalists for the 2025 Ivory Prize at the University of Utah in the policy and regulatory reform category for the approach.
Rountree said it’s been successful so far. The model uses a mix of anonymized hospital, jail, public assistance records and, crucially, prior histories of homelessness to identify roughly 10,000 people who are at the most risk.
That anonymous list is then sent to the county hospital system, where they can identify people and offer them care.
“They call them, they offer them support, they have a conversation, and of course, the program's voluntary,” Rountree said. “People can say no, but when they do say yes, they're immediately connected to a unit that helps them understand why they're in a housing crisis, the nature of that housing crisis, and how rapid cash assistance, in particular, maybe helps stabilize their situation.”
On average, Rountree said, each household helped through the program receives roughly $12,000 in state-funded assistance. But not everyone they reach out to is always receptive.
“It is a cold call,” she said. “Someone gets a call, and a letter, and it says, ‘Hey, do you need help? Can I have a conversation with you about your situation?’ Some people don't trust the government. They don't want to engage.”
Over time, that process has gotten easier with the enrollment rate for assistance improving by nearly 70% since its initial pilot period.
Rountree’s team has had conversations with dozens of local leaders across the country who are curious about how to implement a similar approach. That conversation has yet to take place in Utah.
While incoming state homelessness coordinator and current state Rep. Tyler Clancy has not taken a deep dive into how AI could help homelessness in the state, he is intrigued by it.
“If there's new technology, if there's new, innovative ways we can look at these challenges, we've got to be open to that, so I'm all ears,” he said. “I think my colleagues in the Legislature, same thing. We're focused on results. And so I think that's a great place to start.”
Currently, Los Angeles County uses taxpayer money to fund its predictive AI prevention model. That could be a tougher sell in a Republican supermajority state like Utah that is more willing to cut taxes than raise them.
For Clancy, he would be open to the conversation of having taxpayer money or private philanthropy help pay for a future program.
“For me, it's less about where the money comes from and more about results, making sure that the investment we have is having an impact on the people experiencing homelessness and having an impact on the system,” he said.
Utah has a long track record of philanthropy, including on homelessness. On Dec. 11, Gail Miller and the Larry H. & Gail Miller Family Foundation announced a $10 million gift to maintain and enhance services at the state’s homeless resource centers.
The other good news is that the UCLA model is easily replicated elsewhere in the country. You only need two things, according to Rountree: access to the right data and the political will to act.
“You can avoid that crisis if you get to them in time,” Rountree said. “That's the number one ingredient, the desire to do prevention, and to do it well.”
