Angry Planet - War Robots, AI Targeting in Megacities, and Other Things We Learned From the US Army

Episode Date: June 14, 2019

China and America are two of the world’s great powers. Their economies are intertwined, their military’s powerful, and their soft power spreading across the globe. And tensions are rising. Neither side wants to go to war, but both sides are committed to winning that war should the unthinkable ever occur.Recently, War College’s own Kevin Knodell spent time at the Army’s Joint Warfighting Assessment at Joint Base Lewis-McChord in Washington State. The weeks long training brought together military personnel from seven different countries to train together for a nightmare scenario—a new war in the Pacific.Kevin is the producer of War College, but also a journalist whose work has appeared in Playboy, The Daily Beast, and McClatchy. He’s the co-author of several non-fiction graphic novels—including The ‘Stan and Machete Squad.You can listen to War College on iTunes, Stitcher, Google Play or follow our RSS directly. Our website is warcollegepodcast.com. You can reach us on our Facebook page: https://www.facebook.com/warcollegepodcast/; and on Twitter: @War_College.Support this show http://supporter.acast.com/warcollege. Hosted on Acast. See acast.com/privacy for more information.

Love this podcast. Support this show through the ACAST supporter feature. It's up to you how much you give, and there's no regular commitment. Just click the link in the show description to support now. But yeah, we did talk a little bit about what it would look like if we could program combat AI, being able to give these weapons algorithms and give them presets, kind of templates for what a target looks like. You're listening to War College, a weekly podcast, that brings you the stories from behind the front lines. Here are your hosts. Hello, welcome to War College. I'm Matthew Galt. China and America are two of the world's great powers.

Their economies are intertwined, their militaries powerful, and their soft power spreading across the globe. Intentions are rising. Neither side wants to go to war, but both sides are committed to winning that war should the unthinkable ever occur. Recently, War College's own Kevin O'Dell spent time at the Army's joint warfighting assessment

at Joint Base Lewis McCord in Washington State. The weeks-long training brought together military personnel from seven different countries to train together for a nightmare scenario, a new war in the Pacific. Kevin is here to talk to us about that today. And in addition to being the producer at War College, he's also a journalist whose work has appeared in Playboy, The Daily Beast, and McClatchee. He's also the co-author of several nonfiction graphic novels, including The Stan and Machete Squad. Kevin, thank you so much for joining us. Glad to be here. Okay.

So why is America preparing for a war with China? Well, I think there's a few ways to approach that question. On the one hand, I don't think that, as we said earlier, nobody particularly wants this war. So I don't know that the U.S. military is literally planning for war, but they're planning for the possibility. A lot of that has to come down to China's moves in the Pacific recently. It's become much more, I think, assertive is the polite word that we use when talking about China's more recent actions. It's budding up against its neighbors. It's militarizing islands out in the Pacific.

And it's trying to assert control over shipping routes in ways that we haven't really seen. I suppose that we have seen, but just much more increasing in recent years. Right. They're literally building their own islands in the Pacific. They are literally building their own islands. Okay. So what were the seven countries that participated in this assessment? Well, the United States, obviously, but also the United Kingdom, Canada, New Zealand, Australia, as well as France and Singapore.

Singapore, that's an interesting one. Well, Singapore is actually an interesting one, given that that's going to be where Shanahan is going to be meeting with his Chinese counterpart, I believe. this weekend or sometime within the coming. I don't know when this is actually going to come out. Singapore is interesting also because it is a country with a ethnic Chinese majority, though it has historically been aligned more with the United States and the UK, more so than China. But trade has really been increasing between the two countries.

Okay. And what does something like this look like? Is it a war game? Is it just military personnel coming together and talking? Is it rank and file soldiers? What does it actually look like? How does it take place? Well, the answer is yes, all of the above.

What we were kind of we were looking at there is at Joint Base Lewis McCord, you had officers from all the countries. While across the Cascade Mountains, actually, in the Acoma training center, you had ground troops who were kind of testing out new equipment, but also taking direction from the commanders on the other side of the mountain. There weren't so much rank and file troops from the other countries. It was mostly staff officers who were coming to learn about these systems and learn about what we would do and how they would have to work together better. I know that there had been some discussion about other trips showing up.

I think the Australians and the Brits were going to send. and some ground troops, but it just didn't pan out that way, and it was just the Americans. Okay, so what are some of the threats, especially the technological threats, that each side is preparing for? Well, on the American side, I know that one of the things that they talked about is much more long-range Chinese systems. The Chinese have developed new missiles and recently, apparently tested a rail gun in January. Which is interesting because the United States has been working on the same technology for quite a long time, and we have not entered the testing phase. And that program essentially looks like it's kind of stalling. But the Chinese program, apparently U.S. intelligence believes that that could be part of the Chinese inventory as early as 2025.

The Chinese are also really investing in their cyber capabilities and have been for a long time. They seem to be way ahead of us when it comes to hacking and doing cyber disruption, being able to hack into systems, see intelligence, disrupt systems. That's something that we talked about a lot while we were there. We didn't really get a lot of specifics about what the Americans are trying to do to harden their systems to make it more difficult for the Chinese to compromise them, probably because they didn't. don't want to give the specifics, but it's kind of unclear how we're going about to close that gap. There's this sense, and I'm wondering if you felt it too, that America is lagging behind, especially in cyber capabilities, and that countries like China and Russia are just better at it than we are. Is that something that that was kind of talked about or like in the air at all?

I felt it in the air a little bit, mostly just because I know that to be true. I don't think anybody there wanted to say that or admit how far behind we are. But I think that that's been acknowledged. Now, one thing I would say about that is I'd say that there are obviously a lot of people in the West and in the United States in particular who are very competent with this, but a lot of them are not in the military or working for the military, which is why they've been reaching out so aggressively

to Silicon Valley and tech. That didn't really come up, but that is something that is worth discussing, I think, and worth being aware of. Can we drill down on it just a little bit? Because I think it's an interesting topic, and I think it really plays into current tensions between America and China right now, right?

We have essentially this trade war, on, you know, the executive order that was just signed. Is there a sense, just in your own analysis and reporting, even, that one of the reasons that China is better at this is because they are a totalitarian, you know, government and system, and they are able to move their quote-unquote private sector in the directions that they want it to? And we, you know, the Pentagon has to court Silicon Valley? I mean, that is an interesting question.

I think that there may be some truth to that, though I would say that, and I think we'll get into this a little bit later, that also leads to some problems in the way that they do things. I don't know how adaptive you really are if you are of that totalitarian and that top down. Do you think that then bringing it back around to this assessment that you watched, do you think that that is then a weakness of the child? Chinese military and is the Western side kind of aware of that? And is it an exploitable weakness? I don't know that that really came up. One thing that really did come up is it is a little bit of a mystery what the Chinese capabilities really are. Because, I mean, cyber was definitely a big part of this, but also a big part of this was also how we would actually face them. were there to be an actual confrontation where troops are fighting other troops and tanks are fighting other tanks. And yet you said something you said when we were talking about this before is that you don't see an invasion of mainland China as being on the table, right?

Yeah, well, I just don't see what we would be gaining from it or why we would be doing it. Any confrontation between the United States and China is likely to be somewhere else. perhaps Taiwan perhaps over some speck of oil rich

or important strategic rock out in the Pacific it would probably be on an island it could be

in Asia the mainland but I would sign of itself I don't know why we would be doing that unless things got

really out of control okay so what does what does that what do they think that that looks like then

this tanks and soldiers fighting. And how is it going to be different from any of the wars that we've been fighting recently or in the past? Well, part of it is going to be that cyber stuff. Because one thing that is going to be different is both sides seem to be investing pretty heavily more in drones. Both the aerial drones that we've seen, but also land-based robotic systems and trying to focus a lot more on unmanning the battlefield. field and letting machines do the dirty work, which is interesting, but also I think does present some vulnerabilities when you have everything plugged in and is electronic, which they also did say that they are aware of.

They are aware that hackers or electronic warfare could really disrupt operations, the more we automate things. Right. I think the F-35 is a really good, a really good concrete place to kind of talk about that. Can you speak to some of the vulnerabilities specifically in that system? Yeah, I'd love to. They didn't want to, though. Yeah, the F-35 is connected online all the time through a system called Alice.

It stores targeting data, maintenance data. And it is potentially something that can be used to great effect to make maintenance crews, as well as pilots much more aware about the state. of their aircraft, where they are, what they're doing. But this is all kind of centrally online, and if the Chinese or the Russians were to crack it, that causes some significant problems. They might be able to look in and see what we're doing, or they might be able to disrupt the operations of these things.

And also gets an addition without even them disrupting it. This system software has had a lot of bugs before. Think of any time that you've up in your computer and you had to call IT to fix it or nothing like that. We've had to ground squadrons because they had buggy updates and they had to wait for these things to get fixed until they could get airborne again. So while they can simplify logistics, they can also make logistics much more complicated in that way. It's the Autonomic Logistics Information System. And is my understanding that the F-35 can only fly so long before checking in with one of these systems. So you kind of have to, it kind of has to be grounded at a certain point and, like, log into this thing.

And the system itself is still running and, like, it's not even on Windows 10 yet. They're having trouble getting it upgraded. Um, that's neither, you know, it just, I think it's an interesting place that we're extremely vulnerable that, you know, so what, so you said that they didn't want to talk about these things. Like, what was their reaction when you would ask them these questions? Well, I mean, they were happy to talk about that in a more general sense. But I think the issue came when I, because I didn't point out, um, and they, they, they weren't entirely wrong about this. I compared, uh, the F-35 to the A-10 wardhog. Um, and pointed out that, uh, the A-10, uh, You can't jam it. You can't hack it because it's not plugged in online all the time. And the response was sort of that that may be true, but the A10 doesn't have a lot of the features to protect itself from some of the newer threats in a conventional war. That is very good in a counterinsurgency role against troops that don't have advanced targeting systems, but it might not fare so well in a conventional battlefield. whereas a complex system like the joint strike fighter would.

I mean, they acknowledge that there are vulnerabilities, but their basic premise was that the strengths outweigh the potential vulnerabilities. It kind of leads into my next question. A10 really good for some of the wars we're fighting now, right, against low-tech insurgents. how is the Pentagon adapting the tools and techniques it's learned in Afghanistan in Iraq to the Pacific Theater? What are the lessons learned there and how are they being adapted? Well, from what I could tell, the lesson that they learned is that they don't want to fight wars like that anymore.

And they would just rather not. So I don't know exactly what we were learning other than their mantra is we don't want to, we need to fight the war of tomorrow, not. the last war. So it seems to me more than anything that they just don't want to do that. I'm not sure what all they learned because, yeah, most of the talk was about great power. Let's not think small. Let's think big. Okay, but how do you prepare for the war of tomorrow if you're still fighting the last war? Yeah, that's an interesting question too. I think that is kind of one of the elephants in the room that the war of yesterday isn't the war of yesterday. It's still the war of today, and we don't really know when that's going to end. Yeah, I don't have a real good answer for that.

I think that there's certainly a lot of lessons to be learned from present operations. Okay, let's pivot a little bit. What is Five Eyes and why is it important now and how has it changed over the years? Five Eyes is an intelligence. gathering alliance around the Pacific region. And it's comprised basically of, well, not basically, is comprised of the United States, the United Kingdom, Canada, Australia, and New Zealand. And it was propped up in the 1940s to spy on Japanese movements during World War II. It stuck around during the Cold War to check on whatever the Soviets were doing in the Pacific.

and it has endured. And today it now is used to check on what the Chinese are doing, what North Korea is doing. Now, when we call it an intelligence network, like, do we mean human assets? Like, what does it actually look like? Or do we know? It's more the technology stuff. It's more signal intelligence and also surveillance systems, you know, the eyes, if you will. I think there's a lot of other facets to it.

But basically what it is is it's about intelligence sharing. It's about making sure that if the, like, if the Canadians see something interesting that seems worth sharing that the Americans and the UK and the Australians and the Kiwis also get to know about it. So it's about sharing intelligence and sharing resources to gather intelligence. In Japan has been contributing to it recently, too, right? Recently, to a degree, there's been some cooperation with it, though it needs to be stressed that it's not one of the eyes. It's five eyes, not six eyes yet. Yeah, so Japan is not one of the eyes yet. It is not the sixth eye. We kind of touched on this in a previous episode when we had Kimberly Westenheiser on talking about the Japanese ground self-defense force and Japanese policy that one of the reasons I think, think that Japan hasn't been brought fully into the fold, though there are plenty of advocates for it, is that they've been deemed to be even weaker on cybersecurity than we are. So while they have a lot of assets that can be, they have a lot of knowledge and a lot of great

assets for gathering intelligence, they may not be great at protecting that intelligence, and particularly if they are part of the network as well. Did you see any new technologies? Like, was anybody in robot suits or anything cool? No robot suits, but there were some cool like battlebots, what I would probably call them. They were playing around with what we would call concepts, I think. So they aren't exactly the vehicles that we would probably be seeing once they're actually done, but they're working on them. One of them was a breaching vehicle for taking out minefields, breaking down barriers.

And it was, they basically made it out of the whole of an M1 A1 Abrams tank. But it was part of what they're calling the optionally manned systems. So what it is is it's a, these are vehicles that you can automate a lot of the processes. And this is something that is smart because we talked a little bit about how you can jam a robotic system. You actually can do it manually. You can put troops into it. but you can also take troops out of it and operate them remotely. And commanders can make that decision to plug troops in or take them out.

And it seems like they're doing that by taking older equipment and retrofitting it? Yeah, that was, I mean, in the experimental phase right now, that's exactly what they're doing. Because, yeah, like I said, they're more concepts than even prototypes. And it's mostly old, some of it, fairly old equipment. equipment. They brought out an old M58, which the army really doesn't use very often. Nold tracked a scout vehicle and retrofitted that and put a machine gun on it and used it as it to lay down suppressing fire for the breaching vehicles as they were moving forward. The idea of being that you can just call on other vehicles for support. So you can do the breaching almost automated and also have the support fire be automated. Though actually when we say automated, any distress that these are usually being, we use the term unmanned vehicle, but somebody is controlling these things more often than not.

Remotely operated. Right. Which we can get into why that's a little importance because the military has kind of taken for granted that these things are easy to use. But they actually are fairly complex systems. Well, yeah, let's stick into that then. Why do you think that that's a danger? Well, not necessarily a danger, but something that we need to be cognizant of and aware of. I think when you rely on the robot, and also when we add software and add all these features,

they need to be maintained. And also that the operators can't operate these things indefinitely because it's not a robot. I mean, it is a robot, but it's not. the Terminator. The operator will have to stop or you'll have to trade operators or something. The reason why we should be aware of this is actually some reporting by both of our former boss, David Axe at The Daily Beast. A few years back he reported on how hard the Air Force was pushing its drone workforce. people were working very long hours.

They were keeping pilots operating because it's just a robot. They could keep the operator chugging energy drinks in an air-conditioned room. But these guys were doing it for very, very long hours and sometimes swapping out mid-mission and just keeping these things going. And also the maintenance personnel were very busy all the time. Because we were just using and using and using and using these drones without, with the assumption that because the drones aren't going to get tired, we weren't thinking about how tired the operators were starting to get and how miserable it was starting to get for them. Right. We think it's funny when we've got the, when we have the UAVs in the sky, we tend to we forget that they're being piloted by someone, you know, in a shipping, in an air condition shipping container halfway across the globe. And, you know, we, we forget.

about that human element. And also push the machines too hard, too, like you said. So one of the big buzzwords, we've kind of lighted on this a couple times, is this idea of cyber war. There's been a couple, I have issues with this, because it, like artificial intelligence in the military, I think, we don't have a lot of concrete discussions about what it actually looks like. Along with cyber war, another one of the big buzzwords that I'm hearing a lot, but I don't get

a lot of concrete information on. is artificial intelligence in the military. And you were interesting, we're talking about these drones and they're being piloted, you know, being piloted by people that are being worn out. And it's not exactly a prestigious position

that a lot of people want. Have we toyed with the idea of letting AI drive some of these killing machines? Well, yeah. And we talked a lot about AI and also various software and applications and how they're using them. And some of them are being used in some really cool

and interesting ways. and some of them are being used in ways that I think we need to have longer discussions about. One that I was particularly impressed with was a drone, an aerial drone that the scouts were using out there. And it has some really interesting features and how it's able to feed information to them. It has a bridge classification app that can look at a bridge and evaluate and tell you what kind of bridge it is and roughly how much weight it predicts that it could hold. And it can also do a lot of measurement of terrain features that it used to be that scouts would have to go out with like a ruler and binoculars

and kind of build the lay of the land that way. This drone can basically go out and do that and get all that information for them and then come back to them. And it's plugged into a system that can send that directly to commanders to get like route information of what, looks like and what they're going to be sending people down. That was fairly impressive. We had talked a little bit before we hopped on the air about some auto-targeting systems that they were talking about. Before we talk about, yeah, the targeting systems, I think

what's most realistic and what they're thinking about more in the immediate future is using AI for logistics for simplifying tasks like that. They talked about how they might be able to program routes or program vehicles to be able to understand routes and just move without operators or with a handful of operators. Like the idea that maybe you can command like a squadsworth of cargo trucks and have them be able to figure out how to get in formation and move mostly by themselves. That I think is what they're looking at more in the immediate term. But yeah, we did talk a little bit about what it would look like if we could program combat AI, being able to give these weapons algorithms and give them presets, kind of templates for what a target looks like, and being able to tell them, hey, you know, this area over here shoot at all the things right over here.

That a commander might be able to do that while his troops are engaging one area. They might be able to tell robots to engage another area. Now, they've said, I think the Air Force has said for a long time, that there would always be a human being making the decision to fire the weapon. Is that maybe being backed away from a little bit? So that was interesting because I got different answers depending on who I talked to. In general, that's the idea. What they said is there will always be a human element. They're not going to just completely let the machine do whatever it wants, but they did talk about autonomous targeting.

And they said it pretty at least a few officers talked about that explicitly. One person actually came afterwards and said, well, yeah, but it'll always be up to the person to pull that trigger to approve that target. But then immediately after somebody else said, for now. So they're exploring to what degree this technology could get better. I mean, they are thinking ahead. The technology isn't there yet to the point where they're comfortable with it, but they are thinking that maybe someday it could be. Well, and it also sounds like you're inching ever and ever closer to just letting the robot do it.

If you're talking about essentially using a robot to do all the targeting, to do all the targeting and the robot's going to fire the weapon, but if somebody at some point just kind of comes in and looks at it and says, is okay, yeah, that looks good. I mean, at that point, you're just, you're basically just putting it a human check into the system. And I feel like that's easier at some point, you know, this is just my opinion, to, it's easier at that point to kind of drift into the place where we do have AI that are just doing the combat for us. Yeah, and that was something that I thought about when I was listening to some of the answers. I mean, they're still pretty insistent that that's not going to be the case and they just won't, won't let that happen.

You know, and I don't think anybody has the intention of that necessarily being the case. I don't think that's necessarily what they want. But I couldn't help but think of the opening scene of Robocop where that guy gets killed in the boardroom. Like, I thought about that. Talking about the classic, the good one, not the one that opens with a war in Iran. No, no. Okay. I don't know what film you're referring to.

You just don't acknowledge that the Robocop reboot even happened? What are you talking about, Matthew? There was literally, are you being serious right now? No, I'm not being serious. Michael Keaton was in it. He was actually kind of good, but the movie sucked. Anyway, anyway, God, so many interesting ideas, like, in that movie that they just couldn't quite pull off. Like, Samuel L. Jackson being kind of this reactionary news host and opening with, like, drones conducting a war in Iran.

It just didn't work at all, though, unfortunately. All right, so backing away from that. I want to go out talking about the idea of the battle space for fighting in Asia would be completely different than I think anything that we've, that anything that America has done, possibly ever, I'm talking about fighting in megacities. How do you prepare for that battle space? We talked a little bit about that. So this scenario that they were training for and training in Yakima. There wasn't so much urban that I saw. It was the scenario was it taking place on a Pacific island.

I don't know how populated they were envisioning it being. But I definitely brought that up. And I brought that up specifically in terms of what we were just talking about with these robotic systems. It's hard enough to ask AI to be able to navigate like hills and winding canyons. but asking AI to navigate a densely populated megacity, that's something else even. Because that was actually one of the things that I did ask them specifically was if we were to see, comparing it actually to something that we have done recently, which was the battles of Mosul and Rakhah, which got pretty nasty.

if we had autonomous robots who were picking targets and maybe picking the correct targets, but targets that could be in close proximity to a lot of people or a lot of infrastructure or a lot of property, and these things can just attack, is that something we need to be concerned about? And thankfully, they did say that, yes, that is something that they would think about. And that would also affect whether they would want to use AI in an environment like that. They said maybe we would. Maybe we wouldn't do that. But maybe we would, depending on what intelligence said and what the technology looks like at that point.

But yeah, at this point, certainly nobody would be comfortable with that being the case. I think that that's a horrifying place to go out on, as we kind of like to do. the idea of AI going through megacities and picking targets. Kevin Nodell, thank you so much for coming on War College and leaving the producer's chair and getting in front of the microphone. And you are writing about this for coffee or die, correct? And it's going to be two or three parts. Two parts, I believe. Part one is up, and depending on when this is out, the second part may have come out.

by then as well. One thing I do want to actually clarify, though. They did stress to me that they don't necessarily, I mean, maybe in the future, depending on what it looks like, but they are not advocating for AI combing the streets of megacities looking for targets. When I asked about that, that sounded to them like not something we want. It's something that's something anybody wants. That's all for this week.

you for listening. War College is me, Matthew Galt, Kevin O'Dell, and Derek Gannon. It was created by myself and Jason Fields. Kevin will be away for a while as he travels through the Middle East. We'll be sure to talk to him about what he saw there as soon as he's back. Derek will also be back soon. If you like the show, please like and subscribe. You can find us on iTunes and on Twitter at War underscore College. We'll see you next week. Until then, stay safe.