The NPR Politics Podcast - More Election Stakes: AI Weapons And North Korean Cybercrime

Episode Date: July 8, 2024

Amid a frenetic campaign newscycle, we take a look at some of the very real challenges that the next president will have to tackle during their four-year term including AI-guided weaponry and a complex web of North Korean cybercrime.This episode: national political correspondent Sarah McCammon, cybersecurity correspondent Jenna McClaughlin, and national security correspondent Greg Myre.The podcast is produced by Jeongyoon Han, Casey Morell and Kelli Wessinger. Our intern is Bria Suggs. Our editor is Eric McDaniel. Our executive producer is Muthoni Muturi. Listen to every episode of the NPR Politics Podcast sponsor-free, unlock access to bonus episodes with more from the NPR Politics team, and support public media when you sign up for The NPR Politics Podcast+ at plus.npr.org/politics.Learn more about sponsor message choices: podcastchoices.com/adchoicesNPR Privacy Policy

Hi, this is Lily, sweltering in my ranger uniform. Thank goodness we negotiated for shorts if it's above 90 degrees, but there's only so much that big ol' hat can do. This podcast was recorded at 12.26 p.m. Eastern Time on Monday, July 8th, 2024. Things might have changed by the time you hear this, but I'll still be swearing in junior rangers and hopefully giving really good guided hikes. Enjoy the show. Sounds familiar. Just got to get off to scout camp this week myself. Yeah, sounds like it's hot everywhere.

Hey there, it's the NPR Politics Podcast. I'm Sarah McCammon. I cover the presidential campaign. I'm Jenna McLaughlin. I cover cybersecurity. And I'm Greg Myrie. I cover national security. Today on the podcast, a reminder of the stakes for this presidential election and a look at some of the big challenges the next president will have to confront. We're going to start with AI weaponry. Greg, you've been doing some reporting about weapons being sold by an American defense contractor in Ukraine that are supposed to be able to use artificial intelligence to decide what to target. That sounds really futuristic, but it's not. It's now. So tell us about these drones. Right. Well, so I visited a company called Andral. It's just seven years old. It's based in Costa Mesa, California, just south of Los Angeles. It's one of a number of these new companies that are developing AI weapons. And

their drones in particular are really interesting because you think of a typical drone and the kinds of weapons that have been used. It has an operator guiding that drone, looking at the video footage, sending it places. But of course, there has to be a signal between the drone and the person operating it. But Andral and other companies are making drones that don't require that, that the drone in effect has a brain. So for example, as they're being used in Ukraine, if you want to go take out a Russian tank, you can tell the drone, go in this area, look for a Russian tank, and if you see one, hit it. And the reason is because the Russians can jam those signals and render the drone useless. So Andrel and others are trying to make these drones that can act on their own and literally don't require a human once they've set the target. You know, I have to ask, Greg, I mean, it sort of begs the question,

machines can make errors too, right? I mean, how accurate are these things? Oh, absolutely. You send a drone up there and it's supposed to decide what is a military vehicle and a legitimate target and what is not a civilian vehicle and not a legitimate target. How do you distinguish a soldier from a civilian? You know, is the AI, is the artificial intelligence good enough to make these distinctions? Now, humans can certainly make these mistakes. Now, beyond the technology itself, the brain behind this stuff, the real human brain, the person who founded this company might be familiar to some. His name is Palmer Luckey.

You've been reporting on his background and his current work. Greg, tell me more about him. Right. So I met with him, had a long interview with him when I went out to Andral in California. When he was just a teenager, he built the Oculus Rift. This was the first virtual reality headset for gaming that really hit it big in the consumer market. Just a couple years later, when he was still in his early 20s, he sold it to Facebook for $2 billion. And then he went to work for Facebook, didn't work out.

So he took his fortunes. He decided to get into the defense business. He's a very iconoclastic figure in the tech world. He has a mullet. He always wears shorts and flip-flops and Hawaiian shirts. He's got very conservative politics. He's been a big supporter of Donald Trump for years. So he is not your typical Silicon Valley tech person, and he's certainly not your standard button-down defense contractor. He says his goal is to give the U.S. and its allies the very best tech in the world. It's so good, hopefully it will deter rivals from even fighting. And the better the tech, the better the weapons that can operate on their own, the fewer soldiers you have to put in harm's way. Yeah, so Greg, I mean, when you mention a mullet flip-flops, that's not really what you think of with these traditional

big defense contractors, your Boeings, Lockheed Martins. How has this evolution kind of developed between those guys and this company? Right. So those big companies, you know, they have revenues of tens of billions of dollars. They've been around for decades. They sign a contract with the Pentagon to make fighter jets or submarines or tanks. That contract may last for decades. There's often cost overruns. Stuff is not ready on time. And this is kind of the way defense contracting has been done for decades. But this idea that Andrel has is very, very different. Their idea is to make AI weapons much cheaper, much more quickly,

much more adaptable than these big defense contractors. Some of their drones fold up and can be put in a backpack. So obviously they're not going to carry the same amount of punch as, say, a fighter jet with a big missile on it. But they say this is the future, small, nimble weaponries. And I spoke to Palmer Luckey about why he decided to get into this business. And let's hear what he had to say. I had this belief that the major defense companies in the U.S. didn't have the right talent or the right incentive structure to invest in things like artificial intelligence, autonomy, robotics. And the companies that did have expertise in those things, like Google, like Facebook, like Apple, were refusing to work with the U.S. national security community.

So that's the second part of it, is that there are other tech companies, the big ones he just named, that could do this kind of stuff, but they've really shied away from working in the national security space with the military. And so he saw an opening. Other companies are moving into that space. They're still small. It's still changing, but it does seem to be something we're going to see a lot more of. You know, it's so interesting, Greg. I think everybody recognizes now how much power many of these tech CEOs have, right? We've seen them up in other industries in ways that have big implications for the way that we think about even truth and misinformation. But this is, it feels like maybe another level. We're talking about military power here and sort of upending some of the conventional power players and gatekeepers.

We talked earlier about some of the risks of taking humans out of the equation. I mean, what kind of restraints are in place or are there any? Sarah, there really aren't. There aren't any rules. This is new. It's developing. We don't really have a system to deal with that. Now, some countries and human rights groups, Human Rights Watch in particular, are calling for a formal set of rules about how

AI weapons can and can't be used, how they should be regulated. And, you know, the basic principle, of course, that you just mentioned is you don't want to take humans out of the loop. They should always be part of the decision-making process, even though humans are certainly very fallible. But in battlefield situations, you can see how these restraints easily can break down. It would be very easy to start sending weapons into areas, and maybe the AI is not good enough.

It's not making distinctions that it should be making. And what happens when civilians get hit? Who's responsible? The person who, if you have a situation where a large number of civilians are killed by an AI weapon, is that the responsibility of the soldier in the field who used it? Is that the responsibility of a commander who picked a target or the AI weapons maker who didn't have a system that was sophisticated or subtle enough to distinguish between a military target and a civilian target? We talked about Ukraine. Are we seeing this elsewhere? Are other countries using the same kind of technology?

And how worried are U.S. defense officials about it? Oh, it's a very real thing. Certainly, I think the two big countries are China and Russia. China has very good technology. And if the U.S. and China were to get in a conflict, both of them would be using very advanced AI weapons. Russia may be a little less so, but Russia is very good at things like electronic jamming. So it can block drones and a lot of weapons. And again, I can't stress enough how the Ukraine war has just become a

laboratory. All of these companies doing this kind of cutting edge research and weapons building wanted to get into Ukraine, see how their weapons performed, try to sell it to countries like Ukraine, which desperately needed anything it could get. So what we're seeing is an explosion of these efforts in Ukraine. And wars have a tendency to speed up innovation. And that's exactly what we're seeing here. Yes, Sarah, Greg and I have both reported out of Ukraine, and we're really seeing these technology rules being written in real time. Fascinating. We're going to take a quick break. And when we get back, we'll talk about working remotely from North Korea. And we're back. And Jenna, I'm metaphorically swiveling my chair toward you. You've got a story about an exposed

North Korean campaign to get remote workers hired at U.S. companies and a lot of implications that flow from that campaign. So what's going on with this? Yeah, Sarah, it's a really crazy story. And let me take you back to March 2020. The WHO announces a global pandemic. We all start working from home. And at that moment, North Korea actually saw an opportunity. So its workers could apply for those same remote jobs undercover. So this actually follows in kind of a long line of North Korean scams. They're always trying to make money and they need to find ways to stay afloat to support their nuclear weapons program. And they've done that all sorts of ways. And that includes building stadiums for major sports tournaments. They were selling fake Viagra pills in the 90s. There's

North Korean restaurants in world capitals everywhere. But in the last decade or so, North Korea has really advanced their technical skills because it doesn't take a ton of money or resources to develop those skills. And they work hard, long hours. And Jenna, what kind of scale are we talking about here? I actually spoke to Dr. Jung Pak at the State Department. She's the Deputy Assistant Secretary of the Bureau of East Asian and Pacific Affairs, and she's in charge of diplomatic relations with North Korea, or DPRK. They are generating revenue by getting jobs to develop websites.

They're designing dating apps. The government estimates that there are thousands of workers at dozens of Fortune 500 companies. It's hard to actually pinpoint that thousands number, because when I did some interviews on this, some of these workers are doing more than one job. They're all sort of stuffed together in group houses, you know, 10 plus people in one space. They're working kind of around the clock. And that makes it hard to actually put a number on it. But they are making around $300 million a year. So that's pretty large scale. And this is all happening remotely from North Korea. It's not actually all happening from North Korea. So some of these workers are in places like Russia and China, places that are friendly to Pyongyang, but they're still heavily monitored by North Korean officials. Jenna, I'm always amazed at what the North Koreans

can do in the cyber world. This is not a country with a big tech industry. It seems like, you know, where do they develop these skills? And how have they figured out this huge opportunity to use these hacking skills and intelligence efforts and have so many people who are qualified to do it? Yeah, Greg, that's a good point. And you mentioned sort of they don't have an industry. And that's because they kind of don't need an industry. You know, all of these workers are in service of the regime. So kind of all the resources that they've got go into promoting those ends. It's really interesting what's been happening recently. These IT workers previously were kind of grouped in this bucket of North Korean scams.

But Dr. Jung Pak said that state is now viewing this activity differently. It's not just part of the traditional scam bucket. It's part of North Korea's broader cyber efforts. So when we're thinking about North Korean elite hackers, they do have some similarities with other nation state hackers from China, from Russia. They're breaking into defense industry companies in the U.S. and nearby in South Korea and Japan. They're gathering intelligence from those places. They want to bolster their own defense programs. But, you know, they're also wreaking havoc. Back in 2014, after that movie, The Interview, with Seth Rogen and James Franco

came out, a lot of officials were very angry. They infamously hacked Sony and dumped a ton of sensitive emails from movie executives and movie stars. And, you know, now they're constantly stealing cryptocurrency to fund the regime as well. So a lot of these things are being blended. They're working together. They're getting a lot out of this access that they're having. And Jenna, just to be clear, these jobs we're talking about, they seem like they are just regular IT jobs, right? Yeah, exactly. And, you know, it's interesting how people have tried

to work to uncover what's going on here. I spoke to Michael Barnhart at Google's Mandiant, and he's been studying North Korean hackers for a very long time. And he said, you know, looking back in retrospect, they're able to kind of see this pattern happening, especially once COVID started. So, I mean, what can be done to stop this? Obviously, the U.S. government is concerned about it. I mean, where do these companies that are hiring these workers in or supported by North Korea, where do the companies come in? Yeah, it's difficult. You know, these workers are experts at this at this point. They've had a couple of years to perfect it. And that will take the government stepping in and providing some of this intelligence uh talking to these companies about what to look out for but it'll also take some pretty traditional things like vetting candidates requiring them to

go on video for their interviews uh actually parsing through their resumes calling their references things that will make it really difficult for them to actually fake those things. Obviously, we are dealing with a lot of campaign news right now. You know, Trump was online calling for military tribunals for his political antagonists over the weekend. President Biden is out with a letter to Democrats insisting he will not step away from the Democratic ticket. But as all of that goes on, both of these stories are really reminders of some of the big thorny problems that one of these men very, very likely will have to deal with every day in the coming years. You know, I mean, Greg, I feel like it's just a good reminder of the stakes for this election. It's always struck me that there's this conventional wisdom that

elections are not decided on foreign policy or national security. And yet I always find that kind of strange because presidents have a lot more power to act unilaterally when it comes to these issues. They can more or less start wars or end wars or make big changes to a policy with a country like Russia or China in a way that domestically you have so many constituencies pushing against each other one way or another. A president can't really make these decisions unilaterally. So yeah, I think this is a big issue and we'll be watching. All right, we're going to leave it there for today. I'm Sarah McCammon. I cover the presidential campaign. I'm Jen McLaughlin. I cover cybersecurity. And I'm Greg Meyrie. I cover national security.

Thank you for listening to the NPR Politics Podcast.