Artificial intelligence (AI) is increasingly shaping how transportation systems are designed, planned, and managed in Minnesota. From predicting traffic patterns to analyzing vehicle movements, AI has become a standard tool in transportation research and operations.
A recent CTS webinar highlighted this expanding role of AI in the transportation industry. University of Minnesota researchers Seongjin Choi and Qizhi He, along with Melissa Barnes, operations division AI program manager with the Minnesota Department of Transportation (MnDOT), shared insights on how AI currently supports transportation work and how professionals can prepare for its growing influence.
MnDOT has begun exploring AI through training sessions and small pilot projects, giving employees access to tools such as Microsoft Copilot while fostering opportunities to share knowledge across teams. The agency’s leadership emphasizes that AI is intended to support staff rather than replace them. Barnes said that every AI application is guided by clear principles and human oversight.
AI proves particularly useful in analyzing data from MnDOT’s vast network of cameras and sensors, which generate much more information than staff can manually process. Applications include identifying near-miss crashes, detecting errors in large datasets, and supporting predictive maintenance.
“We see AI as a supplement, not a replacement, for human judgment,” Barnes said. “People who use AI will be outpacing those who don’t.”
To learn more about practical applications, MnDOT surveyed employees, gathering hundreds of ideas that ranged from organizing documents to planning maintenance and analyzing crash data.
The survey found that roughly 60 to 70 percent of respondents felt familiar with AI, about half were excited about it, and 75 percent expressed some level of concern. Barnes agrees with being appropriately cautious.
“It's good for people to be cautious,” Barnes said. “Many people see really high value with artificial intelligence assistance, and we should be responsibly integrating technological advances and not fall behind. Folks asked for even more training and guidelines and hands-on workshops, and [that] really hit on that gradual adoption.”
Choi, an assistant professor of civil, environmental, and geo- engineering and CTS scholar, emphasized that AI literacy is a core skill for emerging engineers. Public and private agencies are piloting AI for operations, maintenance, and planning, and his department has started integrating AI into its curriculum to connect academia and practice.
In Choi’s generative AI course, students explore large language models that support transportation tasks. Through hands-on projects, they use AI to predict travel patterns and detect vehicles and pedestrians.
Rather than treating AI as a “black box,” Choi’s instruction emphasizes critical thinking. He said students are encouraged to review AI outputs and assess when AI is appropriate to use. The goal of adding AI to the curriculum is to ensure it enhances students’ work without replacing their own assessment.
Qizhi He highlighted another aspect of AI in his research on structural engineering. His work on AI-enabled simulation tools for complex material modeling demonstrates that purely data-driven AI has limits. Physics-informed models remain crucial to ensure reliability and accuracy. “AI is not a replacement for engineering judgment or physics-based modeling. It allows us to apply engineering principles at scale and in real time, especially under uncertain conditions,” he said.
—Olivia Hanson, CTS associate editor
