- Artificial intelligence is everywhere.

It's in your pocket, in your home, even at the grocery store.

But what if AI stepped out of our devices and into our streets?

What if it could reshape our cities?

Picture a city where traffic glides effortlessly through intersections, where potential accidents are spotted and prevented before they occur, where flood waters are diverted before they cause damage.

Let's find out how these AI-powered smart cities are putting these solutions to the test, and how they're changing our urban landscapes in real time.

I'm David Hurst.

This is "ncIMPACT."

[upbeat music] When Mark Wittenburg arrived in Raleigh a few years ago from Arizona, he wasn't just stepping into a new role as chief information officer, he was stepping into a city in transformation.

With Raleigh being the third fastest growing city in the nation, the pressure to adapt is constant.

- It is really hard to keep up with everything 'cause it is just moving so quickly.

- [David] But improving a rapidly growing city isn't simple.

For Mark, the first challenge was obvious: traffic.

With each new development comes more vehicles, more intersections, and more potential for accidents.

And there's one question that keeps Wittenburg up at night.

How do you prevent accidents that haven't happened yet?

- We know if an accident happens, we know the state of the intersection, but we don't know if there was, like, a vehicle that nearly missed a pedestrian and then what was happening at the intersection at that point in time.

- [Staff Member] This is 2030.

- [David] For Wittenburg and his team of innovators, the answer isn't just about adding more cameras or hiring more staff.

It's about figuring out how a city can serve its citizens in the age of artificial intelligence.

At the forefront of this initiative are AI-powered cameras strategically placed at key intersections throughout the city.

Using artificial intelligence, they can distinguish between cars, trucks, pedestrians, and cyclists.

This helps engineers spot potential dangers before accidents happen.

- By having these cameras, we can set certain parameters and catch areas of concern before they become a problem or a crash pattern, and maybe make changes so that we can be a little bit more proactive rather than reactive.

- [David] They can use this data to not only make intersections safer, but also make them more efficient, saving time for drivers.

- If we make timing changes and we see a 15-second reduction of delay for a mile or two-mile section and X number of signals, if you apply that to 60,000 people, it can have a monumental impact to Raleigh commuters.

- [David] But Jed thinks there's one more benefit, it could be saving taxpayer money.

- If we can improve the efficiency of a traffic signal, the city wouldn't necessarily have to consider doing a road widening or a capital improvement project, which is significantly more impactful to the people that live along it, costs a lot more, could be a bond.

- Takes a whole lot longer too.

- Correct.

- [David] It was during Hurricane Florence when Kelly Daniel, watching the traffic cameras from the emergency operation center, discovered something unexpected: Those same AI-powered cameras helping manage traffic could be the key to predicting and preventing devastating floods before they happen.

While Raleigh's storm gauges were already monitoring flood levels, combining them with AI analysis of camera feeds offered a new level of predictive capability.

- So if you're waiting till the event happens and you don't have the gauges, you don't have the software to let you know in advance to predict this stuff, and a creek starts flooding over a roadway, well, you know as well as I do, it doesn't take but six inches, maybe a foot at most of water to wash a car down the stream, and you got potential loss of life.

- [Reporter] While Walnut Creek floods, which we're giving you a live look at right now, and spills over- - [David] The system's capabilities were tested in August of 2024 when Tropical Storm Debby approached Raleigh.

Local media reported that Rose Lane, a community with a single access road, became a critical point of concern.

- When we found that was potentially going flood, we went into actions of setting up emergency notifications to be sent prior to the event.

- [David] The data allowed officials to take preemptive action, including adjusting water levels at Lake Johnson.

Water from the lake flows into Walnut Creek, which runs directly under Rose Lane.

- And we did get a little bit of flooding on Rose Lane, but nothing that would stop cars from going going through.

So that was a great thing.

So if we hadn't have lowered it, I'm sure it would've got high enough that you couldn't drive cars through that.

- [David] But as Raleigh embraces artificial intelligence to make streets safer and prevent flooding, I'm curious, where do we draw the line between public safety and privacy?

And how do we know what AI is telling us is accurate?

With 200 cameras watching street corners and AI systems tracking movement patterns, these aren't just theoretical questions.

But Wittenburg says his team has built safeguards into the system's very foundation.

- So at the end of the day, we don't know that you drove through the intersection.

What we know is that a car or a truck or a pedestrian or a bicycle went through the intersection.

So all the data is anonymized.

And really, we're just starting to look at it more from a pattern base.

- How many other cities are using AI?

As Raleigh continues to grow, the pressure to expand these AI systems grows with it.

But for Wittenburg and his team, the human element remains at the center of every decision.

- It's kind of like one of those, do you remember solving those math problems?

And at the end of the day, the professor would always say, "Well, make sure you look at the answer.

And does the answer actually make sense?"

And it's the same thing here, is we wanna make sure that if the AI is telling us that we need to make a change at an intersection, does that change really make sense?

- [David] As cities across America grapple with growth and infrastructure challenges, Raleigh's approach offers a glimpse of what's possible when innovation meets public service.

- So I really believe that as people look at the city services, that they've got an expectation that we're going to be leveraging these technologies to make their life better.

- [David] And the next phase is already taking shape.

A digital twin of Raleigh will soon allow city planners and citizens to visualize how new developments could reshape their neighborhoods, all while AI systems continue learning how to make the city safer and more efficient.

For "ncIMPACT," I'm David Hurst.