E1 | Why Design May Be Transportation’s Biggest AI Opportunity

Jim sits down with John Hibbard, Deputy Chief Engineer at GDOT, to challenge the prevailing narratives around AI adoption in the public sector. They unpack why “buying AI” is the wrong answer, how real value is already emerging in unexpected places, and why the most consequential impact of AI may come not in operations, but in how we design roads in the first place. The conversation confronts uncomfortable truths about vendor hype, accountability, and governance — and makes a clear case that no matter how advanced the tools become, responsibility for our infrastructure decisions must remain with...the humans!
‍

Timestamps:

02:42 – Why AI Is Different From Past Tech Cycles

‍06:07 – Real AI in Action: TMC Incident Support

‍09:59 – Why You Don’t “Buy AI” in Government

‍11:29 – Red Flags When Evaluating AI Vendors

‍14:19 – The Future of AI in Road Design

‍18:18 – First AI Investment for Every DOT

‍20:41 – Governance, Risk & Responsible AI Use

‍24:49 – Leadership Advice: Don’t Be a Bull in a China Shop

‍

Jim: Welcome to At the Intersection. I’m your host, Jim Anderson, and today I’m joined by John Hibbard, the Deputy Chief Engineer for the Georgia Department of Transportation. John also serves as the co-chair of GDOT’s AI Steering Committee and is a great person to provide insight on how a DOT is thinking about AI adoption. Welcome to the show, John.

John: Thanks. Glad to be here.

Jim: All right, John, let’s begin with a lightning round. I want to keep things moving quickly, so I’m going to ask you five questions, and I want five quick responses.

Jim: Number one: what’s the favorite place travel has taken you this year or last year — I guess 2025?

John: Okay, work has taken me to a favorite place. The favorite was a late fall, early winter cruise on the Rhine that we did. That was great.

Jim: The Rhine — all right. How about the title of a good book you read last year?

John: I finally got around to Fahrenheit 451. Excellent.

Jim: At what age did you know you wanted to work in transportation?

John: It took me until my senior year in high school. It was either transportation or music.

Jim: Excellent. That’s its own interesting conversation. What’s your favorite hobby outside of work?

John: It involves music.

Jim: Excellent. And finish this sentence: AI is blank.

John: AI is something that can be very helpful to us.

Jim: Let’s dive into that. You say it can be very helpful. So let’s talk about technology. Obviously, AI is technology. You and I have talked a lot about technology, and you’ve seen technology cycles come and go in transportation. What’s different about AI this time?

John: AI, to me, is a bit different. I have seen cycles come and go, but they were much more tightly focused, at least as I experienced them in my world. AI is broader, and it stands to touch many different aspects of transportation in different ways. To me, that is one of the challenges when I look at AI — where it’s going, whether it’s succeeding, and whatever direction it goes in. It stands to touch all of us in so many different ways, personally and professionally. That makes it more difficult to handicap where it’s going, and much less how well it will do whatever it ends up doing.

Jim: Interesting. So I’m thinking about operations, which is where you spent a good chunk of your career. Technology like traffic signals and intelligent transportation systems — it’s probably an oversimplification to say that’s where the lion’s share of innovation has occurred, but at least a good chunk of it. Designing and building also use technology — construction tech, CAD, and those kinds of tools — but maybe not moving as quickly. Is that a fair inference?

John: I think so. And I say that because, as you noted, I’ve spent so much of my career on the operations side. My knowledge of the design side is more like looking over the fence and musing about how that work is accomplished. My view is a bit different, but it appears to move at a more deliberate pace. The operations side always seems to be impacted by the latest and greatest gadget — although more recently, it’s been less about gadgets and more about data, and now AI.

Jim: Okay, so you’ve helped elevate this from operations to DOT-wide. What would you say DOT leadership broadly is either overestimating or underestimating about AI today?

John: I think many leaders across industries are struggling to assess the reality of AI’s capabilities — what it can do now. It’s hard to put a face on it, so to speak. That sounds odd, because it’s not a literal face, but what can it do that I can understand? How does it help me as an executive DOT leader, or help staff in another context? It’s challenging to make it concrete.

Jim: That leads to my next question: what’s the most useful AI capability you’ve personally seen deployed that’s actually real — not just arm-wavy future talk?

John: Today, I’m seeing the use of an AI tool that helps employees navigate roles, policies, or procedures. Different organizations call these tools different things. In our case, it helps TMC operators when they’re confronted with incidents on the road and need to determine the best response.

Our operators deal with many types of incidents — multi-car crashes, car-versus-truck accidents, even spectacular ones like overturned tractor-trailers. They handle these well because they see them frequently. However, there are unusual wrinkles that don’t happen often. This tool, called TMC React, allows operators to navigate standard operating procedures and receive suggestions, recommendations, or advice on how to respond to a particular incident.

Jim: That’s a very practical example. It makes sense. We often hear AI conversations focus on structured data, high-quality data, and “garbage in, garbage out.” But what you’re describing is unstructured data — words, documents, PDFs. It’s 2026, and instead of flying cars, we’re still using PDFs and Control-F. Everyone listening probably used Control-F this morning. For a TMC operator facing thousands of pages of text, this application is incredibly practical.

John: Exactly. We’re directing people to guidelines that are essentially a pile of text. When something new is added — a rule, a procedure, a situation — it’s not carefully integrated. It’s just added. You get an email saying edit number 2184 was made, and you’re told to govern yourself accordingly.

Jim: Great example of a tactical use case. So let’s talk procurement. You can’t deploy these tools without procuring them, and procurement is challenging in public agencies. What makes AI different from traditional IT or software procurement?

John: It goes back to what I said earlier. AI touches so many aspects of what an agency does. I don’t see someone walking into procurement saying, “I need to buy some AI.” It’s more like, “I need a tool to help identify roadway incidents” or “help manage traffic signals” or “maintain infrastructure.” AI may be part of that tool’s success, but you’re not buying AI itself.

Jim: That’s interesting, especially since every vendor now claims to be an AI company. How do you differentiate between AI vendors who understand transportation and transportation vendors who truly understand AI?

John: I’ve become concerned when a vendor opens with, “Tell me about the challenges you face.” That’s a red flag. It suggests they may not understand the breadth of what a DOT deals with. If instead they reference a specific function, technology, or situation they’ve researched or experienced, that shows preparation and understanding.

Jim: Any other red flags?

John: When a salesperson says, “I have this thing to sell you, and it has AI in it,” I become suspicious. Who defines what AI is? How different is this product from last year’s version, aside from the AI label?

Jim: That makes sense. Let’s look ahead. Five years from now, what part of transportation will be so obviously AI-powered that it feels inevitable?

John: I’ll go out on a limb and say the world of design — designing a road. The design space is already moving in a fundamentally different direction. AI could help evaluate alternatives, identify options, and describe limitations. That world is constantly asked, “Could you do this a little differently here?” AI could help answer that.

I’ll caveat that I haven’t spent a lot of time doing hands-on design. But it’s interesting, because design conversations often focus on digital twins — which are end results. There’s so much work that happens before that: environmental review, right-of-way, permitting, NEPA. AI could help throughout that long process.

Jim: Exactly. The timeline is long, and there are opportunities to integrate data throughout. During construction, information comes in about materials, pipes, asphalt, concrete, signage. Other than checking specs, that information often just sits somewhere. But it could inform operations and maintenance later.

That creates a continuous thread: design, build, operate, maintain. The better that continuity, the more successful you are.

Jim: So if you were advising a DOT commissioner today, what’s the first AI investment you’d recommend?

John: Give staff access to a commonly known LLM and encourage them to use it. Many people don’t know what to do when faced with a blank prompt screen. Access and encouragement are key.

Jim: And that leads to governance. What could go wrong? How do you balance encouragement with risk?

John: Regardless of how a work product is created — pencil, computer, or AI — the responsibility is ours. Engineers tend to be cautious, skeptical, and conservative, which is useful here. AI may assist, but the product remains our responsibility.

Jim: That aligns with the professional engineer model — the stamp, the accountability. It’s a powerful metaphor beyond transportation.

Jim: So looking back, what innovation are you most proud of?

John: In the late 1980s, a systems developer once said, “For years we have built roads aggressively and operated them passively. We need to start operating them as aggressively as we build them.” I think we’re finally getting there.

Jim: That’s a great line. Lastly, what advice do you give younger leaders trying to drive change?

John: Take time to learn how the organization functions before trying to change it. Understand what matters and how the levers work. Don’t be a bull in a china shop.

Jim: That’s a great note to end on. John, thank you so much for joining us.

John: Thanks very much, Jim. I appreciate the chance to talk about where we are and where we’re going.

Jim: Thanks. This is At the Intersection. I’m your host, Jim Anderson, and today we spoke with John Hibbard, Deputy Chief Engineer for the Georgia Department of Transportation. Stay tuned for more conversations with leaders shaping AI adoption across transportation. Thank you.

‍