AI-Powered Cargo Security

The way we picture cargo inspections is outdated and that mismatch is costing us time, money, and security. We sit down with David Smason, founder and former CEO of CargoSeer, now VP AI at Big Bear AI, to unpack what customs enforcement looks like when non-intrusive inspection meets artificial intelligence. Think X-ray and gamma-ray scanning, drones and sensors, and an AI layer that can digest massive streams of images, documents, and historical records then turn it into a decision an operator can trust in seconds.

We get specific about what AI actually does at the border: it doesn’t “replace” the officer, it interprets complexity at scale. David explains how agencies can move from looking at a handful of scenarios to gaming out thousands, flipping the old cat-and-mouse advantage that smugglers rely on. We also talk about identifying exceptions, combining object detection with anomaly detection, and using decades of archived data to catch fraud by comparing what’s supposed to be in a container with what the scan shows.

Just as important, we cover what makes these systems succeed or fail in the real world. Concept of operations, operator buy-in, and workflow design determine whether technology speeds up port throughput or creates new bottlenecks. David shares practical “low-hanging fruit” starting points like empty container verification and automated document extraction, then zooms out to bigger themes: frictionless borders, trust and benchmarking to reduce AI skepticism, smarter government procurement, and why AI should be treated as infrastructure. If you care about supply chain security, trade facilitation, and border technology, this conversation connects the dots.

Subscribe for more conversations like this, share the episode with someone in logistics or policy, and leave a review if it helped. What’s the biggest barrier you see to AI-enabled border security: data, trust, or procurement speed?

Read more

Transcript

Copy Transcript Copied

Meet The Guest And The Mission

Tony Hines 0:13

Well we have a very special guest today, it's David Smason, and he founded and was CEO of CargoSeer, and they carried out non intrusive inspections into cargo using X-rays, gamma rays, and of course artificial intelligence. And that focused attention on which cargo should be inspected physically. And he's going to tell us all about that. So I'm going to leave it to David to fill in the details and tell you how it's happening today. People listening to the episode would probably say well we can imagine all these people at ports and they're going to be inspecting all this cargo that coes in. And they've got an image of people still doing everything and inspecting absolutely every piece of cargo.

David Smason 1:04

But that's not the case at all, is it now? No, sir. With uh with technology, I mean, if you look at the leaps and bounds that technology has made, you know, particularly in the past three years, right? Uh we we all remember or we are all going to remember the day ChatGPT was released, right? I think that's going to be one of our watershed moments in history. Um and the customs agent or the customs operator uh of the present, not even of the future, is this kind of superhuman person, individual, supercharged by technology. I mean, anybody who's gone into one of these modern command and control centers, it's like something out of a James Bond movie, right? You have all these screens, weird machines beeping and bopping and telling you what to do. I mean, I I remember going 15, 20 years ago, right? Uh, and you would have this lone operator going out onto the, I don't, I don't even know what they call it. It's not an apron, it's not a tarmac, it's kind of where they're for the containers. He's kind of walking out there with this old ragged dog, right? I have this image in my mind. This guy like kind of limping out there, the dog is kind of following him, and he's like, sniff the container, and the dog is like, no. Um, but now it's incredible. I mean, uh, you have all these high-tech gadgets, drones, x-ray machines that are doing things that are that are simply incredible, things we never thought we would be able to do. And then we layer on top of that the artificial intelligence layer, right? Which is supporting the operator and taking all of that data in. Because the biggest problem with introducing all of these new sensors and new types of technologies, we still remain human beings, right? The human being, unfortunately, hasn't gotten any smarter, right? The sensors have gotten smarter. But in order to translate what the robots are saying in a scalable manner, we need this artificial intelligence layer. So that's kind of what's happening in the operation center of the present, as I call it, is this one operator is sitting there

The Modern Customs Control Room

David Smason 3:05

supercharged by all of this technology and they're making decisions so fast, so, so fast. Inspect this, contain this, do this, do that, go here, go that. Um, and it's really incredible to see.

Tony Hines 3:16

That's great. That's giving me a different picture of how it is today. I think also, I think one of the things that's interesting is how AI comes into the game. And this is where you really sit, I think, with your interest in AI and the technology behind it. We've got all the hardware, the scanners doing all the work. But the AI is this bridge to intelligence, isn't it? Really?

David Smason 3:39

It is, 100%. And the way I look at AI and the way we look at AI is it's not replacing the human, it's interpreting everything so that the human can make a better decision faster at scale, right? Um, so for all the people who out there who say, you know, AI is going to replace us, or autonomous AI is coming, or our world is going to be run by AI, I don't think so. I think we as human beings, we're always going to be smarter than the computers, right? And it's our job to figure out how to harness this massive power in order to make our lives easier, better, more prosperous, and so on and so forth.

Tony Hines 4:13

Aaron Powell Yeah, it's a tool, isn't it? Really? It's a it it's a piece of technology that we can harness to make better decisions and it improves our quality of decision making because it's giving us, I suppose, more facts in real-time data, things that are happening.

David Smason 4:29

You said that very well. I like that. It's giving us more facts. It's giving us more accurate facts, right? And it's taking into a con. I mean, w when every time I see the back end of our technology, I'm amazed at the different sources we're able to tap into. I mean, billions and billions and billions of pieces of data are all coming together, some of them even in milliseconds, right? And at the end of the day, it's one specific red, green, blue, or yellow for the customs operator, right? Right. So it it's it's really, really incredible. I mean, we're taking in images, we're taking in documents, we're taking external data, internal data, atmospheric data, um, data that was available 50 years ago on specific things, which I can't get in, unfortunately, I can't get into the specific details of that. But you know, think about all the archives that a customs agency has uh has accumulated over the years, we are now able to bring that into the decision-making process of the operator in the here and now. And we found some very, very, very interesting things because criminals don't just appear, criminals just get smarter, right? Yeah. Indeed.

Tony Hines 5:35

Yeah, I mean, that's that's the interesting thing. I mean, how w is this changing the way in which the criminals actually get caught? What you know, what can we actually see that's faster that makes us think, right, I'm gonna target that particular container as opposed to this one over here. That's the one we're gonna take a good look at.

David Smason 5:55

Yeah, I mean, the I that's a really, really good question. Very insightful. I mean, typically or traditionally, the criminal is always one ahead, one step ahead of the cop, right? I mean, from all the old movies, you see they're kind of you know, the cop, the cop is always kind of this goofy guy and the criminal's kind of doing their thing. Um, but we're changing that equation, right? The customs operator, the intelligence officer, whoever it might be who's sitting, you know, um in front of our AI is now able to cover so many vectors of enforcement. It's incredible. Uh, if, you know, traditionally the the law enforcement officer, the customs enforcement officer, whatever it might be, uh, has only been able to look at 10, 15, 20 scenarios, right? If they're well trained and well fed and, you know, whatever it is, right? Um, but now with the assistance of our artificial intelligence, they can look at hundreds, thousands, and sometimes even tens of thousands of different scenarios. So gaming out all these scenarios, and and with criminals and smugglers, I mean, lawbreakers, it's all the same. It's a game of cat and mouse, right? It's like imagine

AI As The Decision Bridge

David Smason 6:55

you have a dam, and law enforcement has to make sure that the entire dam is fully sealed. And if one drop of water gets through, game's over, right? So think about our artificial intelligence solutions and our layer as this superhuman ceiling on that dam, making sure that no matter where the criminal goes, they're not going to be able to get through. And that was previously not not even, it was unthinkable, right? Nobody was even thinking the wildest dreams that would be able to do that. So one customs officer is now able to cover m large sectors, right? And the criminals are zigging and zagging and trying to get through, but they're not really able to because we're able to be there 10 steps ahead of them. That's really the power of the artificial intelligence.

Tony Hines 7:36

Yeah, no place to run, no place to hide. We've got you. Yeah, exactly. Well, that's that's an interesting sort of way to think about what's changing. I suppose one of the things I would say from a port point of view, we've got all this overlay of technology. We've got the AI doing its work, we're picking and targeting what's happening, searching for contraband, perhaps radioactive equipment, perhaps uh weapons that shouldn't be there, and so on.

David Smason 8:07

People that should not be there. It's a big one.

Tony Hines 8:10

Yeah, that's another one, isn't it? People that shouldn't be there. Yes, sir. And I suppose how do we actually ensure that we're not going to slow down the throughput at the port as a result of all the checks that are happening?

David Smason 8:23

That's an excellent question again. I have just one very short answer concept of operations, right? The concept of operations, or in layman's terms, the policy behind the technology needs to be adapted. It needs to be modernized to keep up with this new uh superhuman enforcement, right? And something that we do very well is we understand the technology, we understand what it can do, right? But we also have the deep bench of expertise that allows us to say, we're not gonna sell you technology and not tell you how to use it, right? We're gonna sell you the technology, we're gonna provide the technology to you, but we're also gonna give you the ability to fully operationalize it without impacting your workflows. And that's really where the holy braille is, right? Because you can, today you can buy any technology. You go to the store, you buy a full self-driving Tesla, right? But if you put a two-year-old behind the wheel, they're not gonna know what to do with it, right? So that's kind of uh the analogy that I bring in, right? So we have the best technology in the world, but if the operators don't know how to use it, they don't want to use it, they're resistant to change or whatever, then it's gonna be completely ineffective. And then you're gonna get the effect that you've talked about, the negative effect of slowing down the trade. So what we're doing is we're providing this technology, but we're also providing the user manual on how to do it, on how to use it effectively, right? And that's only possible because we have so much operational expertise. I mean, 20, 30 years back, right, we're the ones who build these workflows, we're the ones who build these SOPs and these operational manuals. So we understand, think about it, we can really go into the mind of the operator and say, this is what you need to do. This is how they, this is how you need to use it. And the operators love it, right? Because it's not somebody from up top coming and saying, from now on, you're gonna use this billion-dollar technology. No matter what, you got to use it. They're not gonna use it, right? Even if you tell them they have to, they're not gonna do it. They might pretend to use it, but then you get the negative effect of slowdown that you talked about, right? But if you come in, you work with them, um, you do it hand in hand with them, then they're far more comfortable with it. And then you're really speeding things up. Um, you're able to inspect 100% of things, you're able to identify where your high risk, high value targets are. It doesn't matter what the specific policy, specific concept of operations is, um, but you really got to do that hand in hand, right? You really got to make sure you're doing both the technology and the concept of operations at the same time.

Tony Hines 10:43

Yeah, I think that's very

Turning Cat And Mouse Around

Tony Hines 10:45

important. In my early life, I learned all about that because I was training people and preparing those manuals, the sort of things that you're talking about to get people to do things and you've got to make them think it's their idea and it's a good one and it helps them to do the job.

David Smason 10:58

Absolutely. So you get it. I get it.

Tony Hines 11:01

Yeah. That's really good. What from a supply chain perspective do you think's changed as a result of these custom checks? If you rethink back down the chain, how's it speeding up operations elsewhere?

David Smason 11:13

Well, you know, it's an interesting question. I I think that if traditionally um supply chain was this convoluted kind of thing that nobody really understands, there's not really any standardization, right? Uh everybody's speaking different languages. I mean, we've both been there. We we know how it is. Um there's two interesting things that are happening. One, customs agencies and governments are are telling the supply chain actors, hey guys, get your game, you know, get get get get your game on, right? Get your stuff in order because we're coming down on you, right? You need to have everything proper, you need to have everything right, you need to get everything filled out. And if there's the slightest discrepancy, we're gonna stop your shipment and send it back, right? It's like getting on a plane without the right visa. You used to be able to do that. Now you can't do it, right? Because the airlines are cracking down. Um, that's the one thing that's happening. So on the government policy side, they're saying, hey, supply chains need to be cleaner. They need to be far more organized and efficient, right? On the other hand, the market is coming in and telling the supply chain actors, all the vendors, suppliers, you know, so on and so forth, hey, we really need you guys to be far more efficient because things are moving so quickly. I mean, if you even look at the subset, uh, microchips, for example, right? How many components do you have in a microchip? I don't even know, right? Could be thousands, right? And they only need to come together. I mean, think about the iPhone in your pocket, right? How many different things needed to happen so that all of the materials got to that specific factory at the right time and were put in the right place so that you could get your iPhone. Now, you as the consumer, you are not forgiving, right? If you don't get the latest iPhone in the store when you want it, you're gonna go over to Samsung, right? You're gonna buy a Galaxy. And this is where things are kind of coming to a head. It's all coming together, and it's these internal and external pressures on the supply chain that are forcing it to change. And I really like that because that's right proving grounds for technology, right? And this is where artificial intelligence technology can really come in and support these processes. And the last thing I want to say is again, it's it's it goes back to our previous discussion on okay, well, how is it changing customer enforcement? We're not trying to affect change externally. We're not trying to come in and say, use all these new brand new systems, change the way you work completely. We're saying, okay, we're gonna come in and support existing processes to make them more efficient and optimized so that everybody wins.

Tony Hines 13:36

Yeah. That's the name of the game, a win-win game for everybody. That's what we want. And I think in supply chains, we can say it can reduce congestion, it can reduce bottlenecks, it doesn't contribute to the bottlenecks. That's the stupidity uh around the inefficiencies that happen everywhere else, I'm guessing.

David Smason 13:54

Yes,

Keeping Ports Moving With Better Ops

David Smason 13:55

sir.

Tony Hines 13:55

And I think one of the things that also struck me, I mean, there's been uh sort of government interventions in various ways. We've got all these tariffs in place now all around the world that we all love or hate. Uh, and uh we we have um lots of people having to do more inspections, I suppose, as a result of those tariffs. Now, if you're on the supplier side, that's great news because that means you've got more work to do, it means you've got more things to provide, more services to uh offer. And I think on the user side, well, we're all going to incur cost, perhaps um, a bit more because we've got to pay the tariffs, but we've also got to maintain those uh registers and checks. Now, it strikes me that where does the cost fall when it comes to all this investment? Is it a government investment that has to happen, or is it the firms that have to pay the cost for all this new AI that's coming in?

David Smason 14:54

Yeah, I I I think that what we're looking at is an initial investment to be shared by everybody, right? The government, the consumers, the vendors. Um but I do think that very quickly we're gonna see it balancing out and the increased speed and efficiency at which things are happening with technology are gonna lower costs significantly, right? So right now we're kind of, you know, think about the height of the apex, right? But as we get lower and as we pass the curve, the learning curve as we call it, yeah, are gonna learn how to procure technology and how to implement technology more effectively. Um, companies are gonna learn how to utilize it more efficiently, and eventually costs are gonna go down, right? Um, without getting into the macroeconomics of the supply chain and the tariffs. Whether the big grocery stores in the US are gonna lower prices on eggs, I don't know, right? No. Or whether gas prices are gonna come down, I don't know. They should, if you look at the economic theory behind it, but you know, the technology is helping everything become more equal, more balanced, and far more quick. And eventually it's gonna be cheaper. I'm a big believer in the fact that chick technology is going to make everything cheaper eventually.

Tony Hines 16:02

Yeah, well, we've got that J curve in operation, haven't we? That you know, where we get the cost going lower as we begin to use it more. And it's shared more equitably. Yeah, I suppose that cost in those circumstances. So I think that's that's an important point that you're making. And I think who knows what the price of fuel or those eggs in the store are gonna be. I have no idea. They change every day. So when we comes to thinking about the borders, are we gonna get to a position where we've got frictionless borders because of AI? 100%.

David Smason 16:39

No doubt. Yeah. No doubt. And and you can see it happening. I mean, everybody's gotten on a plane and had their face scanned and, you know, frictionless boarding. You don't have to pull out your passport uh and you don't have to scan your paper boarding pass anymore, right? That's that's where we're going. The same thing is gonna happen with goods. It's just gonna flow through, it's gonna have everything pre-checked or whatever it's gonna be without getting into the details. And very quickly you're gonna have those eggs on your plate in your house without having to go through all the difficult processes that were previously in place.

Tony Hines 17:11

So I think what I'm hearing is that AI is this intelligent bridge that's probably going to make things smoother in the flow, and we're gonna get fast flow through borders as a result of AI technology being the driving the uh throughput. I suppose.

David Smason 17:28

One 100%. There's no doubt about it.

Tony Hines 17:31

Okay. Now, another thing that struck me, and I was reading about obviously there's kind of uh the fear of China in the background for technology and so on. And there are lots of cases, and there are lots of technology

Supply Chains Get Forced To Clean Up

Tony Hines 17:45

providers coming out of China, as you all well know, and we all know. So is this a threat to governments?

David Smason 17:53

100%. Um the biggest threat right now are state actors who are using means that, you know, um or may not be fair to undermine our national security. Uh and we we as a government and we as private citizens also need to be doing everything we can um to ensure that we have the biggest advantage, the fastest advantage, and and really open up the gate on these. Um and unfortunately, it's taking a little bit of time for people to realize. Um, but if you go around and you see the amount of Chinese technology that is in place, I mean, it it's mind-boggling to me how we have not woken up and said this is the biggest national security issue facing our country today.

Tony Hines 18:35

Absolutely. I think it's a high risk and it's it's a kind of backdoor entry because you can sell somebody the technology and then you can get hold of the data, and then you can actually do things with the data that perhaps you shouldn't be doing with the data. And that's what's happening. And that's what's happening. Uh so you can understand the nervousness of lots of um governments, I think, around the world. Do you think, in terms of the sort of work that you're doing at the moment with your AI, do you actually devise the algorithms? Do you work with teams that make the algorithms that do the uh the smart work in the background?

David Smason 19:14

We work very, very closely with the customer to fine-tune and customize the algorithms for exactly their use case, right? And the really interesting thing about the technology we developed is one of the only customs technologies in the world that is fully customizable, right? Uh and that goes back to my earlier concept, my earlier comment, I'm sorry, about concept of operations. Right. We come in, we understand the exact concept of operations, sometimes even to the operator level, right? Even to the specific crossing at the specific port, and then we take the algorithms and we fine-tune them, right? And we say this is what it's going to do for this specific person, for this specific use case at the specific point in time. And we have all the teams and the technology to do that. Yes.

Tony Hines 19:56

Yeah. And in terms of thinking the flow through, where would you start with a project then? If if you took us through quickly a project, how would how would you start? I don't expect you to give us any secrets here, by the way. I'm just talking about I'm trying to get a picture of the flow through.

David Smason 20:12

You know, yeah, it it all starts with infrastructure. It all starts with data, right? So you say, okay, that this is the data that I have, and this is what I need to achieve. And what are my gaps, right? This is what I need to achieve operationally, and then you can see the gaps, and then you can start building out the infrastructure to get the data that you need. Because obviously artificial intelligence is only as good as the data that you feed it, right? Uh, so that's where we start. We start at basics. We're really, really, really good at coming in and quickly identifying where those gaps are. Like I said, because of our backgrounds in both technology and operations, we can understand that very quickly. And then we'll come in and we'll say, okay, well, once we have the data, let's start fine-tuning these algorithms. Fine-tuning the algorithms does not take that long because of the way that we've built our technology again. And then we start implementing them. And I'm a big believer in two things. I'm a big believer in one low-hanging fruit. So find something easy with a big impact to solve. Two specific examples that I like to bring are empty container verification and documents. When we go into a customs agency, we say, okay, what are your biggest problems? And we see how we can solve the easiest ones with the biggest impact quickly. So empty containers. I don't know if you know this or any of your listeners know this, but today, a container that is empty will get the same amount of attention from a human being as a container that is full loaded, right? And a human being will sit there, they'll analyze it, it'll take five, seven minutes, it'll sit there for an hour and a half, and it's empty for God's sake, right? So you get the algorithm, you say, is there something in there? If there is nothing in there, send it on its way. No reason to stop up the queue.

Frictionless Borders And Security Risks

David Smason 21:50

And the other part is documents, right? There is no reason for a human being to read 50 documents, sometimes in 20 different languages, with handwriting and food, food and literally bug smash. Down then, right? So we come in. So we have the technology that extracts that information. And that's what it's all about. So it's solving those low-hanging fruit problems with massive impact. And then you can start solving the massive problems, right? And then you can start saying, okay, well, how do we find the trigger pin of a nine millimeter Glock hidden in the back cyber container, right? Just as an example. But a lot of people tend to go and say, okay, we're going to do the most complex things first. We believe in approaching those problems, uh in approaching those problems once you've solved the easy ones, because a lot of times the not, it's not really the easy ones. I misspoke on that. The uh more uh, I guess you would say the easier solvable ones, right? With a bigger impact. Um, so once you go in that, and then you can really build out the infrastructure. And then the second thing that I want to say is AI is not a button you turn on. It's not a magic solution. It's a long-term investment, right? You can come in with easy solutions that will give you impact in six months, but it's like anything. It's like a child, right? You can't just have a child and then say, okay, go off into the world and prosper, right? You got to feed it, you got to teach it. You know, that that's how I equate this. Uh, and you know, you're looking at a three-year project, you're looking at a five-year project. Sometimes some customers we've talked with are even looking at this as a 30-year project, which is the right way we all need to be thinking about this.

Tony Hines 23:18

Yeah. Yeah, it is. Uh interesting with the empty containers. There are more empty containers traveling around the world than there are full ones. We all know that. Those are about twice as many, I think. Yeah. And I think and I think with the the easy low-hanging fruit, I think you're talking about Pareto, the 80-20 rule. Let's go, let's go for those things. Are they going to give us the biggest payback in the shortest time? Let's hit those first, and then we can concentrate on the things that are more complex a bit later on. And we might find things out about that complexity as we move through. So I think that's a great uh great analogy that you've given us there, David. I think also when we think about exceptions, I mean that's what we're interested in, isn't it? We're interested in identifying those exceptions when we when we target something. We're thinking, let those empty ones through. If they're empty, that's fine. But if they've got a secret compartment in, as you mentioned, or they've got something in the background that's hidden, then they're the exceptions. And we need to find those exceptions. So we need kit, both the hardware and the software that can identify those. So what can you tell us about that?

David Smason 24:23

Yeah, I mean, there's two ways to approach it, and we use both of them. Way one is look at everything and try to identify the exceptions. Way two is let's only look for the exceptions. Now you can't have one without the other, right? Yeah. First you need to know what's normal, then you need to know what the exceptions are. But the way that the operational workflows today, the way work today, the way that the data flows, you never know, right? It's nice sitting in this nice clean studio and saying, okay, well, if there's a hidden compartment, we'll find it. What if that hidden compartment is is inside of a barrel of concrete, right? We've seen that. Yeah, right. Or under something that's shielding it. Um, that's why you need to be able to take into account both what we call object detection and anomaly detection, right? So we're looking for the anomalies, but we're also looking at the objects and saying, okay, what is the extrapolated anomaly from that? And that's what gives us what we call the 360 coverage, right? A lot of students will look at a specific problem and say, we are going to solve this problem right now, and we're not going to take into account any context. Everything that I've told you throughout our discussion allows us to take into account all of that

Data First Then Quick Wins

David Smason 25:30

context, that 360 context, and say, okay, not only are we going to look for the specific problem in this specific container, we're going to understand where this container came from, where it was before that, who owns it, what was in it before that, right? What the documentation is, and so on and so forth. You get where I'm going with that. Right. So it's all about expanding the scope of your investigation or your inspection. And that was never possible, right? Until we brought AI into this process.

Tony Hines 25:55

Yeah, it absolutely impossible, as you say, because you couldn't get that history and you can't make that history visible in the here and now. And that's what you need to do, isn't it? It's making that archival stuff right up front. What about fraud risk? When we, you know, we we look at exceptions, we think about fraud risk. How can we be better at detecting fraud in the system?

David Smason 26:21

It's exactly what you just said. Look at everything, right? We now have the ability to look at everything. You can look 50 years, 50 years of history, you can even make predictions, right? Uh, we can look at all the commercial data in the world, we can look at all the e-commerce sites, we can look at the entire history of the importer, the exporter, uh, the broker, the agent, everybody who's involved in the transaction, then we take all of that information and we connect it to the here and now, right? Which is the x-ray of the container, and we say, we know what's supposed to be in the container, we know what's actually in the container, let's figure out if they match.

Tony Hines 26:53

Yeah. If we were to look ahead from where we are and say, what developments in AI enabled customs enforcement should business leaders pay attention to over the next five years? What do you what do you think your answer to that might be?

David Smason 27:09

I think it's going to be the distillation of the outputs. So where everybody's trying to get to is to make the output or the decision of the AI more understandable for the for the human, right? Right now we're at a point where it's relatively understandable, and we've done a lot of research and a lot of work on that. And I think we're pretty far ahead of the curve. But where a business leader should look at is how do I make it so that the AI tells the same story that my operators need or my business people need or whatever it is, to understand very, very quickly and without a doubt what the AI is saying. That's number one. And number two is benchmarking, right? We're all heard about hallucinations. Nobody really trusts AI yet, right? Even when you get in your Tesla and you have FSD, full self-driving, you still keep your hands near the wheel, right? Because you never know, right? So I think we're gonna get to a point very soon, you know, three, five years, where we're gonna have the confidence. And like I said, it's gonna start with the easy things. It's gonna start with the empty containers, right? Where it's very easy to say, okay, either there's something in the container or there's not something in the container. But those are the two areas that it it if I was counseling a business leader today, I would say make sure that the AI is understandable and make sure that the AI is trusted.

Tony Hines 28:24

Yeah. Trust is a big thing, isn't it? Do we trust the if we trust the AI, we'd take our hands off the wheel, wouldn't we? Uh and I think that will come, that position. I think also, what would you say to policymakers? What do you think they should be paying more attention to, perhaps than not?

David Smason 28:44

I think that our governments, uh collectively around the world, um need to get better at procure at procuring this type of technology. Um procurement frameworks are built for heavy systems, long-term projects, lots of consultants afford-deployed engineers, and and we do need that, right, to some extent. Um but they're also built for long, long, long processes. And I've been part of so many processes with governments that requirements come out on day one, and by the time the tender is over, 90 days later, the technology landscape has completely changed and what they're asking for is no longer relevant, right? So they need to get better. They need they don't necessarily need to spend more money, right? A lot of people, there's kind of this misconception that that means we're gonna spend more money. We're not, right? We just need to get more efficient uh as a government on the procurement policy and the ability to move fast. Uh, the second thing that I would say

Finding Exceptions And Fighting Fraud

David Smason 29:40

is don't be afraid to fail with AI, right? Um, there's kind of this um mindset with government procurement that, you know, we got to get it right the first time. Uh and there's no going back. I think governments need to build more frameworks for what we call POCs, proofs of concepts, quick, easy, relatively cheap, where you can test almost any type of technology, right, in any type of operational, um operational situation or use case. Uh here in the United States, the United States government is getting very, very, very good at that uh and understanding, okay, we want this technology, we want to test it right before we buy it, and we just need to figure out how we're gonna test it. And that's where the real gap is. So they don't know yet exactly how they're gonna test and how they're gonna measure that outcome, but they're getting much, much, much better at that.

Tony Hines 30:27

Yeah. So smarter procurement is is the way to go, really. It's this compressing time.

David Smason 30:34

It's the only way to go.

Tony Hines 30:35

It's the only way to go. Yeah. I'd agree with you. I think it is, it's compression. So what do we think, if you're advising government today, your top priority, what would it be when it comes to customs technology and border security? What would you say to them?

David Smason 30:53

I would say invest in infrastructure to bring the data, to understand what the data is, um, and to make the data AI ready, because that's the foundation. It's like, I mean, we've seen a lot of governments, a lot of procurement processes, they'll go ahead and they'll buy technology, and I equate it to buying a house without a foundation, right? It looks nice, but the second you move your family and the whole thing starts sinking, right? So think long term, think about infrastructure, think about data, think about concepts as operations, take a couple months to mull it over, right? And then start buying things and start testing in everything that I previously said.

Tony Hines 31:30

Do you think that they should think about AI as infrastructure?

David Smason 31:35

100%, no doubt.

Tony Hines 31:37

Yeah. Because it is really I know it isn't a bridge, it's not a port, but it is important. And perhaps even the most important.

David Smason 31:45

I would fully agree with that.

Tony Hines 31:47

Yeah. David, we've had a quite a rounded conversation about the things that are involved here, and you've given us some good insights into how it works. I suppose what I ought to do is just ask you over to you and say, what do you think you haven't told us that you think you'd like to tell us that's really important?

David Smason 32:08

Well, that's a great question. Um Well, I've I've talked a lot, probably a little bit more than you, right? I mean, just get very passionate about this. And they're yeah, you know, it's it's interesting. I I I would encourage our audience to talk about this, to discuss it, to have open conversations, to not be afraid to talk about the difficult things, right? Um, and more than that, to understand that AI is not a magic button, right? There's no potion that you can pour in and say this is gonna happen overnight. Um, we need to understand that we are at the cusp of a change, right? We're already deep in it. Uh and our lives are gonna look completely different in 10 years. And the more we talk about it, the more proactive we are about this thing that's called AI, the better off we're gonna be. Uh, and we all need to come together.

Trustworthy AI And Smarter Procurement

David Smason 32:55

Um, I think that the second thing that I would like to say is we need a lot of cooperation, international cooperation between agencies, between governments, between people, right? Talking about how we can use this because this is not something that's gonna be limited to one country. It's not something that's gonna be limited to one type of people. It's gonna going to affect all of us.

Tony Hines 33:13

I think those are really good points, and that's a good position where to leave it. So, David, thanks very much for coming along today and talking to us on Chain Reaction. I hope it's gonna go great with your new projects and everything you're doing.

David Smason 33:28

Thank you very much. I appreciate it. Thanks very much indeed. Thank you so much, Tony.

Tony Hines 33:34

Chain Reaction. Well, there we have it. It was wonderful to talk to David Smason because he's the man doing the job and he knows what's going on. And it's always good to hear firsthand about the challenges and the opportunities that face us. I hope you've enjoyed the episode as much as I've enjoyed the episode talking to David. And we thank him for his time and his knowledge and experience. Well, that's it

Final Takeaways And Sign-Off

Tony Hines 34:05

for this week. I look forward to seeing you on Chain Reaction in the future. But for now, I'm Tony Hines, I'm signing off, and I'll see you next time. Bye for now. Stay informed, stay ahead. Listen, learn, lead with chain reaction.