A panel at the 2025 Utah Transportation Conference in Salt Lake City on the future of transportation urged state and local agencies to move from experimentation to practical, near-term deployments of artificial intelligence and related sensing technologies.

Sean Wilson, former secretary of the Louisiana Department of Transportation and the panel moderator, framed the session by saying transportation is a bipartisan, results-focused field: "We believe in getting things done for the people that we serve, regardless of who they are or where they live." He introduced private-sector and civic leaders who described pilots and applications ranging from lane-level sensing to digital delivery in construction.

Why it matters: panelists argued that AI and better sensing can improve safety and operational efficiency now, not only in a distant future of autonomous vehicles. Erin Mendenhall, mayor of Salt Lake City, said the promise of the technology is tied to saving lives: "The bottom line has to be about saving lives," she said, citing 27 traffic-related fatalities in Salt Lake City in 2023. Speakers added that making AI work for transportation will require higher-quality data, workforce training and a willingness to deploy rather than plan indefinitely.

Panel highlights

Tyler Duvall, chief executive officer of Cavinu, said his company aims for "full roadway coverage, lane level accuracy" by combining machine vision and edge computing so agencies can detect hazards and debris more precisely and act faster. He said such systems can be deployed in months and are aimed at reducing both fatalities and recurring congestion.

Jim Anderson, CEO of Beacon Software, cautioned that "AI is not one thing, it's many things," and urged transportation professionals to understand differences among machine learning, large language models and generative AI. Anderson recommended giving AI multi-paragraph, project-level assignments (such as synthesizing large document sets) rather than restricting its use to minor editing tasks.

Construction and quality control

Mike McArthur of the Ralph L. Wadsworth construction team and Chris Peterson of Horox (formerly a UDOT employee) described digital delivery, 3-D modeling and lidar scanning as concrete ways to improve construction quality. McArthur said those visualizations give crews and reviewers a shared picture of what must be built and can speed verification that work was completed correctly. Peterson summarized the workforce challenge in blunt terms, quoting a technology officer: "AI is going to make smart engineers smarter and dumb engineers dumber," and added that organizations must lean into new-hire skills rather than rely solely on legacy ways of working.

Data, governance and deployment

Multiple panelists raised data quality as an immediate bottleneck. "Garbage in, garbage out," attendees heard; improving workflows and data consistency was described as a prerequisite to reliable AI outputs. Panelists urged agencies to pilot deployments that solve operational problems (signal optimization, curb management, incident response) rather than indefinite planning: "Try. Just try," one speaker said, recommending practical experiments over purely academic pilots.

City use cases and equity considerations

Mayor Mendenhall described pilots using LIDAR for signal optimization and sketched an example curb-management system that could reserve parking and extend meters based on real-time detection, while acknowledging that some residents prefer simple, low-tech parking options. Panelists said leaders should design deployments that preserve access for older or less-tech-savvy users.

What the panel asked agencies to do next

Panelists urged four near-term steps: improve the quality and interoperability of data; define operational goals (not just technical vision); run real deployments that solve specific problems; and enable — rather than overly restrict — staff experiments. Several panelists stressed that investments may require up-front money but will yield safety and affordability benefits over time.

No formal actions or votes were taken during the session. The panel concluded with a call to tell the story of successful deployments to build political and financial support for larger investments.