AI in waste management: An ever-changing solution

Tech executives have a clear picture of the ways artificial intelligence (AI) can be used to accelerate safety and efficiency in waste and recycling collections, it’s just a matter of staying current with the rapidly evolving technology.

“It has moved so fast, and I can’t keep up with the number of AI applications that are out there,” said Brian Sano, head of strategy at Intuitive AI, a software developer based in British Columbia. “There’s 20 new ones every single day.”

Sano was among the panelists in the session AI-Powered Waste Solutions: A New Era of Efficiency and Safety at WasteExpo 2025 in Las Vegas. He was joined by Stevan Mikha, chief technology officer at Prairie Robotics; Tamey McIntosh, president of Global Sensor Systems Inc.; and Samantha Podgorny, digital product manager for Environmental Solutions Group for a discussion about the integration of AI in waste management and recycling, ranging from the development of solutions for automated braking systems to 360-degree camera systems and reverse automatic braking.

AI solutions

Panelists spoke about the importance of company culture and prioritizing safety to ensure the success of AI-enabled solutions.

Safety might not be a revenue-generating tool, but it’s a cost avoidance measure, and Podgorny said she sees potential to leverage AI to create a safer environment for workers and drivers.

Across the industry, cameras, driver monitoring systems and advanced driver assistance systems are being implemented to improve safety during collection. Environmental Solutions Group, based in Chattanooga, Tennessee, has led the way with its Soft-Pak software and 3rd Eye 360-degree camera systems, offering a bird’s-eye view of what’s around the truck.

The company’s AI-powered driver monitoring system, or DMS, uses internal-facing cameras to monitor the driver and scan for things like drowsiness, smoking, failure to wear a seat belt or looking at a cellphone, she explained.

Advanced driver assistance systems use external cameras for object detection, looking for nearby people or bikes, and monitoring lane segmentation to ensure drivers aren’t crossing over a solid line.

Taking cues from cars

Currently, Podgorny said, the average car on the road is much more technologically advanced than the average waste collection truck, even though driving a collection truck is exponentially more dangerous.

“From a technology standpoint, we’re trying to figure out, OK, what are people getting used to in their cars that can translate over to a garbage truck?” she said. “Unfortunately, garbage is super nuanced, and what should be easy is difficult.”

While automobile manufacturers can build the same car model over and over again at scale, waste collection trucks are typically tailored to each specific use case, which contributes to that difficulty.

“With garbage trucks, every single one of them is a little bit different, so it’s a lot harder to scale than Nissan or Toyota,” Podgorny said. “But that’s what we dream of—how can we make that garbage truck just as safe as my car? There’s a fine balance of active safety without too many bells and whistles and noises that are distracting. That’s something we think about every day.”

Environmental Solutions Group continually examines what’s being implemented in passenger vehicles, especially around safety, to see what could be tweaked to work in waste collection. Later this year, the company plans to upgrade its 360-degree camera system for trucks to include object detection and reduce blind spots for drivers.

Another innovation Podgorny is particularly excited about from Nissan is its Invisible Hood, a camera that shows what’s underneath the front part of the truck.

“That’s just an area that I hadn’t thought of before,” she said. “It was like, ‘Oh, how can we translate this?’ especially for [commercial] front load … or a residential front loader—things like that.”

Shifting demand

Global Sensor Systems, an Ontario-based truck sensor technology company, offers infrared sensors for reverse automatic braking. An upcoming partnership with Environmental Solutions Group will integrate a data component to capture media and event information when drivers apply the brakes.

To meet demand from customers in bigger cities, McIntosh said, Global Sensor Systems is expanding its offering to provide additional information on braking events to decrease the frequency of unnecessary braking.

“We’re shifting our product focus … for those who want something more robust that can go down an alley without setting off the sensor,” McIntosh said. “When it sees the wall, it’ll know that it’s an alley wall and it won’t set off the sensor.”

For Environmental Solutions Group, AI’s biggest safety impact has been around using data to change driver behaviors and habits, Podgorny says. The company is focusing on predictive analysis to better understand driver behavior.

For instance, if a company were to notice that a driver had far more speeding instances on Tuesdays and Thursdays than any other days of the week, it might use that information to reprimand the driver for speeding. However, digging into the data could unearth additional factors affecting his or her behavior that might not have been obvious.

Perhaps that information shows unbalanced routes, with more stops on Tuesdays and Thursdays than on other days of the week. Rebalancing the routes could result in reducing speeding instances dramatically.

“I think it’s important when you’re getting a lot of information faster, don’t just jump to conclusions and look at the wrong picture,” Podgorny said. “Because if you use too much of a stick without the carrot, drivers can be your loudest complainers” rather than a hauling company’s best advocates, she added.

Keeping pace

Sano stressed the need for continuous technology updates to keep up with AI’s rapid evolution.

To address the issue of frequent tech churn, Prairie Robotics, Saskatchewan, has been working to build a data backbone for automated model deployments, Mikha said.

“For us, we deploy regularized hardware, but we know that it’s going to change in three years when you get new trucks, whatever that looks like,” he said. “We’re building a data backbone, so it doesn’t matter what piece of hardware we’re plugging in on the trucks.”

On top of integrating regularized hardware, Mikha discussed the importance of providing automated model deployment that is designed to help future-proof operations.

“We get data in and don’t need to worry about this fleet or that fleet, we’ll just retrain automatically on the first of every month and deploy that,” he said of Prairie Robotics’ approach. “I think that’s where things are going to move because that’s the kind of scale that’s required.”

Panelists agreed that it is important to gather clean data and use predictive analytics to improve safety and efficiency in waste hauling operations, with McIntosh emphasizing that good data ensures future readiness and adaptability.

“Things are changing so rapidly, but I think the thing that all of us can do is just contribute data [that is] as clean as possible,” she said, “so no matter what happens in the future, you’re ready.”

Despite the integration of AI, humans remain essential in waste management and recycling operations. For instance, every time a reverse automatic braking system sounds an alarm, the system captures a photo or video and creates an event that is sent to a cloud-based portal.

“The data is great, but if no one is there to actually look at it or enforce it, then there’s no point,” Podgorny said.

The panelists, each of whom is focused on implementing AI-enabled solutions to improve safety in the waste and environmental services industry, said the ultimate goal is to help the industry get out of its stubborn position on the Bureau of Labor Statistics’ list of the top 10 most dangerous jobs.

“We all need to make sure that safety is a priority,” Podgorny said. “We [also] want to be more efficient. Hopefully, AI solutions can help with that.”

The author is the former managing editor of Waste Today.