UVU expert: AI in public safety may not be a net benefit.

Is the rapid adoption of artificial intelligence in public safety actually making communities safer — or simply introducing new risks that outweigh the rewards?

A Utah Valley University expert argues the answer is far from clear.

Brandon Amacher, who directs UVU's Emerging Tech Policy Program, says the breakneck pace of AI development in emergency services is forcing a difficult reckoning. "These conversations are extremely nuanced," he said, "because just because a new capability exists doesn't necessarily mean that it is a net benefit to the public good."

The stakes are playing out in real time in Weber County, where an AI system now handles roughly 10% of non-emergency dispatch calls.

Weber 911 Executive Director Devin Rose has watched the technology mature at a startling rate. "[When] we started looking at this, it wasn't there yet," Rose said. "A year and a half later, it's night and day difference."

The pitch behind the system is straightforward: by letting AI field routine non-emergency calls, human dispatchers can devote more attention to genuine emergencies.

Rose's ultimate ambition is for AI to handle 100% of those calls, a target he says could be reached sooner than many expect.

The public is paying attention — and pushing back

Not everyone shares that enthusiasm. When 2News showed the AI dispatcher to Soren Puckett, his reaction was blunt: "Sounds kind of dystopian."

That sentiment extends beyond dispatch lines.

Flock camera technology, which captures license plate data across jurisdictions, has drawn similar skepticism from residents wary of pervasive surveillance.

Amacher acknowledges that emergency management and law enforcement often operate with limited resources, which makes AI's promise of efficiency genuinely appealing — "if properly implemented," he added.

Trust, not technology, is the real obstacle

Our analysis suggests the biggest challenge facing AI in public safety isn't technical capability but public perception. Amacher put it starkly: "That is really the core of the problem.

It's not how often it's being abused. It's how people feel about that relationship with government being reformed oftentimes without their consent."

This points to a three-way tug-of-war between industry leaders pushing innovation, policymakers trying to regulate it, and a public that often feels left out of the conversation entirely.

The result is a paradox: the more sophisticated AI becomes — Amacher describes the trajectory as going "from very, very rudimentary to extremely capable" — the more unsettling it can feel to the people it's meant to serve.

According to Amacher, discussions about secure and responsible use are happening, but they remain highly nuanced.

The central question, he suggests, is not whether the technology works, but whether its adoption happens with public awareness and consent.

What should communities do now?

For everyday residents, the most practical step is to stay informed and demand transparency from local agencies before new systems go live.

Public meetings, records requests, and direct questions to elected officials can help ensure that government adoption of AI happens with consent, not in the shadows.

For smaller agencies considering AI tools, the lesson is to communicate openly about what the technology can and cannot do — and to be honest about its limitations.

Building trust early is far easier than repairing it after a misstep.

Amacher said no one truly knows what comes next, which is part of what makes this moment so uncertain.

But he emphasized that decision-makers who are upfront about their choices will have a far better chance of earning the public's confidence.

Whether AI in public safety becomes a net benefit may ultimately depend less on how smart the technology gets — and more on how honestly it is introduced.

Original reporting by KUTV 2News, Salt Lake City. Quotes and data drawn from the KUTV article "Trust may be the biggest hurdle as AI moves fast in public safety."