The ongoing introduction of Artificial Intelligence (AI) into fleet operations will not only redefine the industry, but also benefit customers, drivers, and companies big and small, according to a panel of AI experts at this year's ATA conference in Nashville, Tennessee.
The panel included:
- Jim Burg, President and CEO of James Burg Trucking Company
- Rusty Kirby, Director of Transportation at Pilot
- Steve Schmidt, Senior Community Health Specialist, Drive Safe Missoula
- Evan Welbourne, Vice President, artificial intelligence and data, Samsara
Before diving into the complex subject that is AI, moderator Jim Burg asked the panel what, exactly, is not AI.
"AI is not an autonomous machine that's able to do things its own way," Schmidt explained. "Right now, it's in a generative format, meaning it requires a human intelligence input asking it to do something. Only then can it generate a response or output."
A couple of popular examples currently in operation are ChatGPT and Google Gemini.
"Because AI is so broad, it's a technology that enables machines to do things humans normally do, such as language and generating new text. It's not just one thing but rather an all-encompassing technology, ranging from the cloud to games on your watch," Welbourne added. "AI won't automate away all of the grunt work. In the long term, it'll probably manifest with all new types of services and products we can offer customers. It will create new jobs."
Where AI could become overwhelming for any business is how the data organized. "The challenge is organizing it and making it useful," Kirby said.
From left: Jim Burg, President and CEO of James Burg Trucking Company; Rusty Kirby, Director of Transportation at Pilot; Steve Schmidt, Senior Community Health Specialist, Drive Safe Missoula; and Evan Welbourne, Vice President, artificial intelligence and data, SamsaraClean Trucking
Where's the value?
It's important to remember that AI's introduction into the trucking industry is relatively new and many fleets are still struggling to understand why it's worth the investment. However, AI is not a brand new technology. In fact, it's been around since the 1950s.
One example of its value is generating reports, analyzing data, and responding to emails. Human oversight, however, is still necessary.
"Some of the flashy consumer electronics like robot vacuums [already use AI]. But really it's all over the place, like in your thermostat and the smart grid behind your thermostat," Welbourne said. "The biggest area for direct interaction is probably a search engine like Google. It's already trying to complete the query as you type. It's trying to help you get to the result you want. The results are not just the classic search engine results but rather a summarization answer to your question at the top of the page.
AI's first commercial adapters were tech companies that used it to optimize their supply chains. The banking and healthcare industries were next in the adoption process.
"Then the dot com boom came in the 90s," Welbourne pointed out. "The latest wave of generative AI came in the past few years. Adoption now is different. It's all happening very fast now. For trucking, the difference this time is rooted in the [Covid 19] pandemic. Some businesses were slowing down during tough conditions. Operations and logistics leaders had to kick things into overdrive to keep the economy running. Turning to technology, for companies, was a solution. Post-pandemic, AI hasn't stopped. Trucking is now there with the early adopters, and that normally wasn't the case. Many fleets are already using AI and many will double down over the next year. Trucking is now at the adoption forefront."
[Related: Volvo introduces AI-powered preventive maintenance for trucks]
Reshaping physical operations
What's for certain is that AI will not replace drivers, at least not in the near future. However, the technology can benefit well-qualified drivers in many ways.
"What's changed is our ability to optimize around trucking," Kirby said. "There's a revolution around telematics, GPS; how do you flip new tractors to a new supply source? Where are they? What are they delivering? This creates a lot of opportunities. It's a huge volume of data and predictive dispatching can help. A lot of input goes into that."
Utilizing AI to help manage data and materials will get the job done faster, Schmidt believes. "It can free up some our mental time and energy. Something can now become a two-person job, instead of, say, fifteen."
Fleets are also seeing safety improvements thanks to AI. For example, the Samsara driver app has helped save half a million lives a year, according to Welbourne. Denver, for example, "has seen a big impact by using safety products - a 94% reduction in accidents. DHL has seen a decrease in accidents overall by 26%. Accidents that still happen are now costing 49% less. We also released a new drowsiness protection feature. One company had one drowsy incident per day. It's now down to 0 in past 30 days."
Getting AI up and running
Any company and/or fleet wishing to get into the AI field but isn't sure how need to know the integration process begins with something critical: data.
"It's best to start with an audit of your data. Where can you plug in AI for benefit? The quality of the data is important, so it's important to be accurate with this," Welbourne said. "You also want to have some expertise, whether in-house or with a partner. Give it a test drive for a few months."
Cameras are another type of AI that's already available. "They're looking for specific behaviors and are very customizable," Kirby explained. They're now focusing on things that cause accidents, like speeding."
Deployment risks
Deploying AI does not come without risks, but even those risks can provide valuable lessons, especially financial ones. Take McDonalds' drive-thru ordering kiosks, for example. The fast food giant recently announced it is removing them.
According to Schmidt, "they were initially thought of as job killers, but they actually created more work in the kitchen."
Customers started ordering more food because of the ease in which the kiosks functioned. However, this required more kitchen staff to fulfill that need.
"What was initially thought of as a job killer, but rather, in this case, a job transferer. Also, AI will always ask for that upsell. If someone fails to ask for that upsell for whatever reason(s), AI will always ask in order to generate more sales. They're always there on task," he added. "We view AI as an assistant to the worker. The privacy of the worker is just as important as the safety. We're empowering workers. The system is very configurable. It could, for example, warn a driver in the cab if they're not wearing their seat belt or if can't put their phone down. It's a lot about offering the controls that makes it accessible for anybody."
Data, security, privacy issues
Another common concern regarding data are issues surrounding security and privacy. Will AI create more or less problems?
"It's all about control and transparency over the features, such as retention settings or uploading data to the cloud or not," Welbourne said. "AI can become like a black box; how does it come up with decisions? How do we find a way to open the box for transparency while also keeping the human in the loop? That's a big part of it."
It's important to know that AI running on a company's internal systems are not necessarily dependent on internet access, meaning it's important to keep everything all on one server behind a firewall for security.
"Managing data properly. How is the data stored for the driver? Explain why you're doing it for a better experience. Make goals achievable," Kirby stressed.
Rules and regulations
As the use of AI continues, it's important for lawmakers to regulate it. But it's also vital for consumers and companies to understand how it works.
"As long as people are educated about what AI is capable of doing right now at this point, [we're in a good position]," Schmidt asserts. "You can only generate something that results from a human intelligence input. Obviously, some form of regulation is needed."
Achieving a proper balance between regulation and advancing the technology is vital.
"We need a balance between what you regulate and over regulate," Kirby said. "What makes sense is to be pretty broad with guidelines. The way things are changing you need to pre-regulate. Transparency and control is the level we have to get to and avoid bills that are overly prescriptive."
The AI race
AI rollout in other countries, particularly China and Russia, will have an effect as to how things will play out in the U.S.
"I think a lot about regulation but it could be dangerous to think about it too much," Welbourne says. "When there's an AI race, a lot is how we embrace it. How we take advantage of it. There a lot of processes that need to be integrated."
According to Schmidt, AI is currently in the generative phase, but it'll be quite different in the near future.
"It's going to become what's called 'agents,' where agents will be able to actually do things," he said. "[For example, you can] ask a Google agent to book a trip, make sure you have a window seat, a reservation at this hotel and/or restaurant, such as the one I ate at six months ago but forgot what it's called. That's the next phase and it's probably two to four years down the road, maybe sooner."
Small fleet investment & new jobs
Small trucking fleets have a lot to gain by incorporating AI but it's equally important to know how to properly invest in it.
"There are more tasks geared towards efficiency," Kirby said. "Predictive failures and things OEMs can leverage to help correct issues in tractors. We're going to have to rely on vendors for this based on the amount of development and computing power."
Learning from AI is important and questions that need to be answered include "what can I do with this data? What are the options," Welbourne said.
The technology will undoubtedly create new trucking jobs and expanded roles in existing ones, such as information technology (IT).
"IT will have a lot of AI familiarity," Kirby acknowledged. "This is a good application for your fleet [IT guys will be able to contribute]. Here's how you use it."
Welbourne's conclusion is, perhaps, the was the biggest takeaway from the discussion: "Being able to partner with AI, whatever form it takes - that'll be the new form of job. That could manifest with questions about your data and operations. That could be a new category of job."
