Pothole ahead: AI, the new eyes on US roads

As America’s aging roads fall further behind on much-needed repairs, cities and states are turning to artificial intelligence to spot the worst hazards and decide which repairs need to be done first.

Hawaii officials are distributing 1,000 dashboard cameras as they try to reverse a recent surge in traffic fatalities.

The cameras will use artificial intelligence to automate the inspection of guardrails, road markings and pavement markings, and instantly distinguish between minor problems and emergencies that require dispatching a maintenance crew.

“This isn’t something they’re going to look at once a month and then sit down and figure out where to put their pickup truck,” said Richard Browning, Nextbase’s chief commercial officer.

Hawaii drivers will be able to sign up for a free dashcam worth US$499 ($A766) over the next few weeks as part of the Eyes on the Road campaign, which was piloted on shuttles in 2021 before being paused due to bushfires.

The state faces unique challenges in maintaining legacy highway infrastructure, said Roger Chen, a University of Hawaii engineering professor who helped facilitate the program.

“Equipment needs to be sent to the island,” Chen said.

“They have a space constraint and a topography constraint to deal with, so it’s not an easy problem.”

The companies behind the technology specifically boast of its ability to detect damaged guardrails, although the program also monitors things like street debris and faded paint on lane markings.

“They analyze every guardrail in their state every day,” said Mark Pittman, CEO of Blyncsy, which combines dashboard feeds with mapping software to analyze road conditions.

Hawaii transportation officials are well aware of the risks that can arise from broken guardrails.

Last year, the state reached a US$3.9 million ($6.0 million) settlement with the family of a driver who died in 2020 after striking a guardrail that was damaged but never repaired in a crash 18 months earlier.

In October, Hawaii recorded its 106th traffic fatality of 2025; This number is more than all deaths in 2024.

It’s unclear how many of the deaths were related to road problems, but Chen said the grim trend underscores the timeliness of the dashboard program.

After San Jose, California, began mounting cameras on street sweepers, city staff confirmed that the system accurately detected potholes 97 percent of the time.

Now they are expanding the effort to park inspection vehicles.

But Mayor Matt Mahan said the effort would be much more effective if cities added their images to a common artificial intelligence database.

Mahan said the system can recognize a road problem it has encountered before, even if it occurred elsewhere.

“What he sees is, ‘Oh, this is actually a cardboard box sandwiched between two parked vehicles, and that counts as roadway debris,’” Mahan said.

“We can wait five years for this to happen here, or maybe it’s right at our fingertips.”

San Jose officials helped launch the GovAI Coalition, which went public in March 2024, for governments to share best practices and eventually data.

The state of Colorado is a member, as are other local governments in California, Minnesota, Oregon, Texas and Washington.

Not all AI approaches to improving road safety require cameras.

Cambridge, Massachusetts-based Mobile Telematics has launched a system called StreetVision that uses cell phone data to identify risky driving behavior.

The company is working with state departments of transportation to determine where specific road conditions fuel these hazards.

Ryan McMahon, the company’s senior vice president of strategy and corporate development, was attending a conference in Washington, D.C., when he noticed that StreetVision software was showing multiple vehicles braking aggressively on a nearby road.

The reason was that a bush was blocking the stop sign, which the drivers did not see until the last second.

“What we’re looking at is an accumulation of events,” McMahon said.

“This brought me to an infrastructure problem, and the solution to the infrastructure problem was a pair of secateurs.”

Texas authorities are using StreetVision and various other artificial intelligence tools to address security concerns.

This approach proved particularly useful recently when 402,000 kilometers of roads were scanned to identify old street signs that were long past due for replacement.

“If something was installed 10 or 15 years ago and the work order is on paper, God help you find it,” said Jim Markham, who handles crash data for the Texas Department of Transportation.

“Having an AI that can scan and scan this is basically a force multiplier that allows us to look wider and farther, much faster than just walking around.”

Experts in AI-based road safety techniques say what’s being done now is largely just a stepping stone to a time when the majority of vehicles on the road will be driverless.

Blyncsy CEO Pittman, who worked on the Hawaii dashcam program, predicts that within eight years, nearly every new vehicle, driverd or driverless, will come with a camera.

“How do we see our roads today from the perspective of a grandmother in a Buick, but also Elon and his Tesla?” said Pittman.

“This is a really important nuance for transportation departments and city agencies. They’re now building infrastructure for both humans and autonomous drivers, and they need to start closing that gap.”