The use of drones is transforming how the war in Ukraine is being fought. Squadrons of pilots, mostly young men, are deploying skills and reflexes learned from videogames to attack Russian troops and inflicting casualties, while acting like military startups in the field. WSJ’s Ukraine bureau chief, James Marson, joins host James Rundle to discuss how this new crop of nontraditional soldiers are developing the future of modern warfare. Plus, we put artificial-intelligence chatbots through their paces, with a weekend in the woods.
Sign up for the WSJ's free Technology newsletter.
Learn more about your ad choices. Visit megaphone.fm/adchoices
[00:00:00] Brauchen wir wirklich noch einen Computer? Alle wahrscheinlich nicht. Aber wenn du Musik mit der Power eines Neuralprozessors neu erfindest oder unterwegs Migrationsmuster mit einem ganztägigen Akku analysierst oder deine Ideen mit dem KI-gesteuerten Co-Creator zum Leben erwächst, dann kann ein Co-Pilot Plus PC einen Unterschied machen. Nicht alle brauchen einen leistungsstärkeren KI-Computer. Aber wenn du versuchst, die Welt zu verändern, auch wenn es nur deine eigene ist, haben wir einen für dich entwickelt. Microsoft Co-Pilot Plus PC mit Snapdragon. Die bisher schnellsten und intelligentesten Windows-PCs. Die Akkulaufzeit variiert hier nach Nutzung und Einstellungen.
[00:00:33] Welcome to Tech News Briefing. It's Thursday, November the 14th. I'm James Rundle for The Wall Street Journal.
[00:00:40] Chatbots powered by artificial intelligence have exploded in popularity. But how well can they truly replicate human friendships?
[00:00:47] We put four bots from the biggest tech companies through their bases. And then drones are changing the battlefield in Ukraine and even the nature of military units as tech savvy pilots target Russian tanks and infantry.
[00:00:59] We look at how one squadron is operating like a tech startup on the battlefield.
[00:01:07] But first, AI chatbots have become startlingly realistic, driven by the emergence of generative AI and large language models like OpenAI's ChatGPT.
[00:01:17] Tech companies are pitching the bots as your helper and even your friend. But is their humanity just a paper-thin veneer?
[00:01:23] To find out, WSJ's senior personal tech columnist Joanna Stern spent 24 hours at a cabin in the woods with four chatbots made by Meta, Microsoft, Google, and OpenAI.
[00:01:36] The results offer a glimpse into the technology's potential and its limitations.
[00:01:40] Joanna joins us now to talk about her weekend away with AI.
[00:01:44] Joanna, what are some of the interesting things you discovered about interacting with the bots over 24 hours?
[00:01:49] So I was really trying to test for two things. One was friendship and the other was assistants or helpers because we've all used Siri and Alexa and we all know that they've hit a wall.
[00:02:00] You can set timers, you can play music, you can do some of these very clear tasks.
[00:02:05] If you want to go a step further, it's typically it doesn't understand you or it's giving you a web result and it can't do the thing.
[00:02:13] But these are very different. They're very conversational. They're based on the large language models, so they kind of talk about anything you want.
[00:02:20] They never don't have an answer. So I designed a few challenges. One was to build a fire together because the woods and, you know, we're at camp.
[00:02:27] I asked these bots, we've got to build an outdoor fire. How do we do it? And they all pretty much guided me the right way.
[00:02:33] You'll need tinder, which catches easily.
[00:02:35] Kindling and fuel wood.
[00:02:37] Meta specifically helped me figure out how to use a hatchet.
[00:02:40] Hey, Meta, how do I use a hatchet to take off a branch?
[00:02:43] To remove a branch with a hatchet, hold the hatchet at a 45 degree angle, swing it into the branch near the truck and chop through the wood in small controlled strokes.
[00:02:53] They also do a really good job of helping you cook. I wanted to make some meatballs for dinner.
[00:02:58] They guide you through that. Not as well step by step, but very good in terms of what are the ingredients you need? How do you start?
[00:03:05] But when it came to the friend test, as you said, these are based on large language models.
[00:03:10] There are certain limitations that come with that in terms of memory, in terms of just the limits of a AI model interacting with human emotions.
[00:03:18] Can you talk about some of that?
[00:03:19] One thing you just noticed from listening to these voices is how friendly they are.
[00:03:24] Almost annoyingly friendly.
[00:03:26] Everyone should take a listen here to what Microsoft Co-Pilot sounds like.
[00:03:30] How are the vibes this afternoon? Just good vibes. Peaceful vibes. Forest vibes. Cozy vibes. Authentic vibe.
[00:03:36] It says vibes all the time. And yes, it made me feel more comfortable.
[00:03:40] And you feel like, oh, this is really friendly. We're really getting along.
[00:03:43] And Meta lets you set the voices to some celebrities.
[00:03:47] I chose Kristen Bell.
[00:03:49] I can help you with things like creating images, answering your questions or giving you advice.
[00:03:54] And she sounds so friendly.
[00:03:56] You feel like, oh, I'm like this is she's my real friend.
[00:04:00] And that's where you realize this is just a voice.
[00:04:03] These words are just sort of empty.
[00:04:07] Because when I started to really get into sort of the things you'd want to do with your friends, right?
[00:04:12] Laugh.
[00:04:13] They kind of laugh, but they don't really know what they're laughing at.
[00:04:17] I mean, there was this amazing moment.
[00:04:18] We're sitting by the fire and ChatGPT flubs a sentence.
[00:04:23] And this is what happened.
[00:04:25] Tell me how the fire making you feel.
[00:04:27] How the fire making you feel.
[00:04:29] Fire make me feel good.
[00:04:31] That's the spirit.
[00:04:33] Like there's an emptiness in that laugh.
[00:04:36] Another place is emotion.
[00:04:37] These bots don't really have emotion.
[00:04:40] They even say, I'm a bot.
[00:04:43] I can't experience emotion.
[00:04:44] I'm made of code.
[00:04:45] And so that kind of stinks when you want to talk about something a little bit more in depth.
[00:04:51] And then the lack of memory is also just comical.
[00:04:55] But again, makes you realize like this is what human friends are for.
[00:04:59] These will all get better memory.
[00:05:01] But will the emotion ever be there?
[00:05:04] I don't know.
[00:05:05] Well, I think this is a really interesting point that you raise in your story as well.
[00:05:08] Are there concerns that as this technology develops, people may lose sight of the fact that these are effectively user interfaces for complex math equations, right?
[00:05:16] That's what's going on with the hood.
[00:05:18] That's exactly right.
[00:05:19] These voices and these personalities, it's just disguised for the fact that these are computers.
[00:05:25] There's computer code that runs in the background.
[00:05:28] I hit that wall pretty quickly with them.
[00:05:30] I knew that there wasn't much depth to these relationships.
[00:05:33] But my fear isn't about people like me.
[00:05:37] I'm pretty well adjusted.
[00:05:39] I've got a few friends.
[00:05:40] I've got a really great, caring family.
[00:05:43] But some people don't have that.
[00:05:45] They're lonely.
[00:05:46] And especially younger folks who are coming to these platforms or looking to social media, et cetera, to form connections.
[00:05:55] And so these types of apps are right there.
[00:05:57] And now there's a voice in them.
[00:05:59] And we've seen an example of this.
[00:06:01] Just recently, a teenager struggling with mental health issues had been chatting with a bot on character AI.
[00:06:07] And he took his life.
[00:06:08] And the company did update its safety guidelines after that.
[00:06:13] But these text windows or the microphones to talk to these bots at any time we want is there.
[00:06:19] That was our senior personal tech columnist, Joanna Stern.
[00:06:23] Coming up, drone warfare in Ukraine is providing a window into how wars will be fought in the future.
[00:06:29] But who are the people piloting the greatest threat in today's battle space?
[00:06:33] That's after the break.
[00:06:34] The future of warfare is being defined on the battlefields of Ukraine right now.
[00:06:50] Drones are reshaping how the war is fought in profound ways while a new class of soldier is emerging in the conflict.
[00:06:56] In fact, they don't resemble traditional soldiers or military units at all.
[00:07:00] WSJ Ukraine Bureau Chief James Marson tells us about the rise of the drone pilot as a dominant force on the battlefield in Eastern Europe.
[00:07:07] And how they're using tech to hold back the Russian advance.
[00:07:11] James joins us now to explain.
[00:07:13] So James, how are drones changing the battlefield in Ukraine right now?
[00:07:17] Initially, they were used for reconnaissance.
[00:07:19] But as time went by, the Ukrainians also started to use them for strikes as well.
[00:07:23] So initially, they were using 3D printed claws, which they were able to put hand grenades or other explosives in.
[00:07:31] Then the next step from there was that Ukrainians took what's known as first person view drones.
[00:07:36] So these were racing drones.
[00:07:37] They're about the size of a dinner plate.
[00:07:39] And they were previously used by racing.
[00:07:41] The pilot wears goggles and can see a feed from the camera in the goggles, controls it via remote controller.
[00:07:47] And they attached explosives to these drones.
[00:07:49] They cost about $500 to make.
[00:07:52] And with the explosives attached, these will slam into a target and explode.
[00:07:56] So the Ukrainians have tended to be ahead of the Russians, ahead of the curve.
[00:08:01] They're a small country.
[00:08:01] They're a very creative country.
[00:08:03] They've got very good engineers.
[00:08:04] And they've been a little bit ahead of the Russians in terms of the developments.
[00:08:07] But the Russians have obviously been adapting as well.
[00:08:11] Catching up, copying some of the technology, using some of their own technology and ideas.
[00:08:17] So the Russians are never far behind.
[00:08:18] And what the Russians are able to do is also produce it at big scale because they're a bigger and richer country.
[00:08:24] So you've spoken to a number of drone pilots in Ukraine.
[00:08:28] Typically speaking, what are their ages, their backgrounds?
[00:08:31] What's the profile of these pilots?
[00:08:32] So most of the drone pilots tend to be young guys.
[00:08:36] They're often in their early 20s.
[00:08:38] They're not typical people you would think of as high-caliber soldiers, special forces or something like that.
[00:08:44] You have this kind of movie image of people as muscle-bound and able to run for a long distance and take extreme pain and et cetera.
[00:08:52] Most of these guys, you wouldn't take a second look at them if they're walking along the street.
[00:08:56] And a lot of them are avid computer gamers.
[00:08:58] They're skilled with their fingers.
[00:09:00] They're skilled at analyzing videos and seeing where they need to send a drone.
[00:09:04] So the skills are a little different from what you'd expect from perhaps a traditional soldier.
[00:09:09] But obviously there are some very strong similarities.
[00:09:13] The aim of these drones is to kill, to stop the Russians who have invaded Ukraine.
[00:09:16] You have these young guys who are killing huge numbers of Russians because of the effectiveness of this weapon.
[00:09:24] Chris Kyle, a very famous American sniper, is reputed to have killed about 150 people.
[00:09:30] You know, I spoke with guys who themselves said, I'm not one of the best.
[00:09:34] You know, I'm pretty good, but I'm not one of the best.
[00:09:35] And they'd killed 300 Russians.
[00:09:37] So that speaks to the effectiveness of this weapon, which also, of course, has a psychological impact on these guys.
[00:09:42] You know, they're young guys.
[00:09:43] Some of them use humor to try to not think about it.
[00:09:47] But at the end of the day, a lot of them say, well, this is the situation we've been put in.
[00:09:52] We have to kill these Russians to prevent them from taking over our homeland.
[00:09:56] And so we're prepared to do it.
[00:09:57] And the psychological consequences that may come a little later, we'll deal with that when we have to.
[00:10:02] But this is something we have to do in order to defend our homes.
[00:10:05] Quite a remarkable burden to put on young men who are in their 20s with that level of violence, even if it's out of a move, right?
[00:10:12] It must be pretty affecting for them.
[00:10:15] Yes, it's pretty horrific.
[00:10:16] I've seen a lot of the videos and they publish a lot of the videos.
[00:10:19] You know, you can see them on social media.
[00:10:21] You're seeing bodies being blown into pieces.
[00:10:24] It's very graphic.
[00:10:25] As you say, there is a kind of removal, a distance from it because you're seeing it on a screen.
[00:10:31] But still, you are doing that to another human being.
[00:10:35] So it is, you know, it's not the same as perhaps thrusting a knife into somebody.
[00:10:39] But it's certainly not the same as if you're piloting a bomber and dropping bombs.
[00:10:44] You're not seeing who you're killing.
[00:10:46] You're seeing who you're killing.
[00:10:47] You mentioned earlier the non-traditional backgrounds of these soldiers.
[00:10:51] And that's also manifested itself in the character of the units as well.
[00:10:54] These drone squadrons don't necessarily resemble traditional infantry units as much as they do creative startup-like culture sometimes, you reported in your story.
[00:11:03] Very much so, because the key element here is that Ukraine has had to get creative to take on the Russians.
[00:11:09] Russia's population is about four times the size of Ukraine.
[00:11:13] Its military is much larger, much stronger.
[00:11:16] So they've had to get very creative to take the Russians on.
[00:11:19] Part of that has been through the culture that they've created here.
[00:11:23] You've got a lot of companies that have come out of nowhere.
[00:11:27] They've been created during the war and are now producing thousands of drones.
[00:11:31] The Ukrainian government has reduced regulation.
[00:11:34] It's cut taxes.
[00:11:35] And companies have responded.
[00:11:37] You have situations where these companies have links with military units.
[00:11:40] So they'll test out drones.
[00:11:41] They'll send them to the unit.
[00:11:43] The unit will be able to test them out.
[00:11:44] Then you've got charity funds that, you know, if those drones are effective, the charity fund will then buy those drones in bulk and send even more of them to that unit, to other units.
[00:11:54] So it's a really fast capitalist war economy, essentially, that Ukraine has fostered.
[00:12:00] And it's one of the things that's keeping Ukraine in the fight right now.
[00:12:03] There is one important point here.
[00:12:05] Drones aren't all powerful.
[00:12:06] Russia has developed and has put on the battlefield electronic jammers, which are able to overload the signal between the controller that the pilot's using and the drone itself, which then causes the drone to crash.
[00:12:20] So actually, only about one in three, at best, of these strike drones are hitting their targets.
[00:12:27] In the future of warfare, we're going to see a lot more of this kind of technological competition.
[00:12:33] That was our Ukraine Bureau Chief, James Marson.
[00:12:36] And that's it for Tech News Briefing.
[00:12:38] Today's show was produced by Julie Chiang with supervising producer Catherine Millsop.
[00:12:42] I'm James Rundle for The Wall Street Journal.
[00:12:45] We'll be back this afternoon with TMB Tech Minute.
[00:12:48] Thanks for listening.

