On this episode of the AI For All Podcast, George Matus, Founder and CEO of Teal Drones, joins Ryan Chacon and Neil Sahota to discuss AI drones and AI in warfare. They talk about the history of drones, the current and potential applications of drones, drone swarms, the ethics of AI drones, drone depiction in popular culture, US-China relations, and the future of drones.

Always passionate about flight, George Matus had his beginnings in remote control modeling. He became a test pilot for an RC manufacturer at age 12 and after building a number of custom drones in high school, George started Teal in 2015 to create a new era of flight.

Since the inception of Teal Drones, George raised several rounds of venture capital funding, launched four products to market with American manufacturing, and continues to serve as CEO of the company after its acquisition by Red Cat Holdings. George is a Peter Thiel Fellow, Forbes 30 Under 30 member, has competed on the national television show BattleBots, and holds over a dozen granted patents.

Teal Drones, a subsidiary of Red Cat, is focused on rebuilding America's drone industrial base. The company's overall mission is to accelerate the adoption of drones around the world, and specifically for defense, to give warfighters superhuman capabilities. Teal’s current flagship product is designed to "Dominate the Night" and is being deployed at scale across government and commercial segments around the world.

(00:35) Introduction to George Matus and Teal Drones
(01:09) History of drones
(02:34) Different types of drones
(03:44) Role of AI in drones
(04:48) Military vs consumer drones
(07:43) Why drones are getting so much attention
(09:31) The value of AI drones
(11:26) The holy grail of drones
(13:15) Engineering challenges of AI drones
(15:19) Potential of drones
(16:57) Drone swarms
(19:45) Ethics of AI drones
(22:19) Western vs Eastern views on technology
(23:26) Drones in popular culture
(25:33) US-China relations
(29:02) Future of drones

- [Ryan] Welcome everybody to another episode of the AI For All Podcast. I'm Ryan Chacon, and with me as always is my co-host, Neil Sahota. He is the AI Advisor to the UN and Founder of AI for Good. Neil, how's it going on your end?

- [Neil] I'm doing all right. Just another day in paradise.

- [Ryan] And we also have Nikolai, our producer. Nikolai, how's it going?

- [Nikolai] It's going good. I'm looking forward to this episode because I am a certified drone pilot.

- [Ryan] We are going to be focusing on drones, applications of AI in drones for a variety of different use cases, talking about what the future might hold when it comes to the evolution of AI and drones together.

And to discuss this, we have George Matus, the Founder and CEO of Teal Drones, which is a subsidiary of Red Cat. Their mission is to really focus on accelerating the adoption of drones around the world, and specifically for defense, to give warfighters superhuman capabilities, which is very interesting, and I definitely want to dive into that.

But first, George, thanks for being here.

- [George] Thanks for having me. I'm excited to spend the next 30 minutes or so diving into drones and AI and the future of that. It's a little hard for me to do that with my wife, so this is gonna be great.

- [Ryan] Absolutely. Let's kick this off. Let me ask you a high level question just to get our audience up to speed on things.

Where do things stand currently and where have we come from when it comes to the drone industry? Like, where are we? What does the history look like? Just to give people kind of full scope before we dive into more specifics.

- [George] Yeah, that's a great question. I'll try to keep things pretty high level and quick because there's so much to talk about, but the drone industry is really interesting, and I think extremely unique in that it's gone through some big inflection points over the last eight years. In 2015, I would categorize that as like the peak of the Gartner curve hype cycle, where tons of people were excited about drones, a bunch of venture capital funds were investing money into drones.

People were expecting that by 2020, drones would be everywhere doing so many different things. But in 2017, we reached our trough of disillusionment. We saw a lot of consolidation in the industry. And we saw China and DJI monopolize the drone market and the drone industry. Since then, we've been on our path to enlightenment, quote unquote, and I think we're starting to hit some really positive inflection points for regulation, for technology. AI is a big part of that, and I'm really optimistic about our ability to help rebuild America's drone industrial base and show the world the incredible potential that drones have.

- [Ryan] One thing I wanted to ask you, to tack on there, is when we're thinking about drones, maybe you could lay out the range of products and markets that drones exist in and what they do. Obviously, I think a lot of us have seen military applications of drones being used as well as personal drones, right?

So there's a wide range of when we talk about drones. Maybe just summarize it up for audience of how they should be thinking about the drone industry as far as what they're being used for and the difference between kind of the different varieties of them.

- [George] Yeah. So to be clear, we're talking about small drones.

Small unmanned aerial systems. These are typically like three pounds or less. They can be used for consumer applications for commercial public safety and defense applications as well. Most people are familiar with DJI. They build awesome flying cameras that have dominated the market and have been used by basically everybody around the world.

But there's some issues with that that we'll talk about later and which also is part of why Teal exists.

- [Ryan] How is AI playing a role in the drone space? What features does AI bring to drones? And how can, how should we be thinking about them together?

- [George] Also a big and fascinating question. Just as with every industry, AI is about to have a massive impact that I think a lot of folks are underestimating.

With respect to drones and more specifically with respect to Teal, we are very focused right now on the defense segment. So our drones are being used for intelligence, surveillance, and reconnaissance. And what we're trying to do, as you mentioned earlier, is give warfighters superhuman capabilities, better protect them, build an asymmetric advantage for our country and deter any conflict in the first place.

So AI has a huge impact in that across a number of things, which I'm sure we'll dive into here soon. But I think it's enabling a really big inflection point for drone technology that will finally enable drones to be unlocked in their potential.

- [Nikolai] What would you say sets apart these military drones from consumer drones?

Because something that I find remarkable is that a lot of these military drones are actually pretty similar to consumer drones and actually even quite cheap. Like we've seen, I mean, we're seeing even DJI drones being used on the battlefield. So, it's interesting how accessible these drones are, especially for use as reconnaissance or even weapons to basically any nation. So, what's your thoughts on that?

- [George] Yeah, I think to level set sort of the impact of drones in defense, we're hearing from all of our sources and based on what's happening in Ukraine, that the invention of the drone is as impactful as the invention of the machine gun in the late 1800s.

It has incredible potential in many different ways. There is actually a lot of convergence across requirements for consumer drones and military drones. They have to fly for a certain amount of time at a certain range with a certain kind of camera. Teal actually started off as a consumer drone company back in 2015.

We were building consumer drones that were primarily differentiated because of their modularity and because they had a really powerful compute platform on board to be able to run AI at the edge. We've taken a lot of that experience from the consumer market and applied it to our drones in the military market since we pivoted to that segment as of about 2019 and so all that's to say that a lot of the fundamentals are similar, but there are a few very distinct differences.

So with respect to military drones and what we're seeing now in Ukraine is that it's very easy to jam and interfere with a drone. Either in the data link, which it's being controlled by, by a user, as well as with its GPS. So if you don't have a data link and if the drone doesn't have GPS, it's rendered useless, and you can't do anything with it.

I think a big focus right now in this military segment is being able to create resilient autonomy, which is powered by AI, ultimately to enable drones to overcome that sort of contested environment and still be able to fulfill the quote unquote kill chain. The kill chain is essentially the process under which a military works.

It's composed of three parts, which is to understand, decide, and execute a task. And basically a military drone that's enabled with AI allows those three things to be done a lot better than a consumer drone.

- [Nikolai] It seems drones have been around for a little while, but we're only now hearing about them really taking off in use in war and on the battlefield.

Like, why now is it really taken off?

- [George] I think it's for a few reasons. There's a lot of industry dynamics that I started mentioning earlier. With respect to China and the US, but I'll put that on the side for now. I think the biggest reasons are because of missed expectations for regulation and technology.

So again, in like 2015, people thought that by 2020, drones would be everywhere doing amazing things in all these different areas. But it didn't happen. So some of the big inflection points that we're going through now is that a few years ago it took multiple people to fly a single drone for a commercial operation.

For the last three years until now, one person can fly a drone for commercial operations or military operations. But the inflection point that we're hitting as we speak is the ability for a single person to control a whole swarm of drones, a whole fleet of systems that could not be possible without increased autonomy and AI.

So that's one big technological inflection point that we're hitting. There's a bunch of others that have more subtleties to them, but I think that's a really notable one. And then on the regulation side specifically, in America, domestically, we're starting to unlock the ability for drones to fly beyond visual line of sight over different sorts of areas and over people which inherently unlocks the number of use cases that those drones can have.

- [Neil] George, are you saying then the utility is directly tied to the manual component of drones, that if a person can now actually control a swarm, it's, I don't know what the right word for here is, there's enough value getting created? It's cost efficient? What's the connection?

- [George] Yeah, so if you think of a single person being able to intelligently control a whole fleet of drones, that reduces, and using AI, that reduces the cognitive workload for that operator, so many new things become possible.

I think there's some really positive use cases to mention here that we're part of outside of defense, which is wildfire and land management. So some of our big customers right now are the U.S. Forest Service and Cal Fire in California. Imagine having a fleet of drones flying over and covering very efficiently a large amount of land and with AI and the high resolution multispectral sensors on board the aircraft, you can very quickly map out an existing wildfire or predict where a new one is going to start. Search and rescue, imagine having 20 drones flying up the side of a mountain and with those same technologies be able to very quickly find a missing person.

You can quickly start to see where things are going and how crazy things are about to get.

- [Neil] I think those are actually a couple of really good examples, I think, for our audience. As a guy that's done a lot of work with AI and wildfires, I think most people don't realize the traditional way we would try to detect wildfires is people literally looking at an area with binoculars.

And that's obviously not that efficient. Maybe not that effective either, so I think that's a great example. I got to make my cultural reference here. If you've watched the movie Spider-Man: Far From Home, and they're able to control that swarm of drones with those glasses, is that what we're talking about? Is that kind of the holy grail we're trying to get towards?

- [George] It is and that might be a good segue to quickly talk about where AI will live and what it will be used for. In the context of defense, we talked about the kill chain, being able to understand what's going on, decide what to do about it, and then execute a certain task.

Some of those fundamental things also transition over and apply to other use cases in in the enterprise and in public safety worlds. And then the AI can live not only on the drone, but also in the human machine interface and and the controller that an operator uses to control the drone. So the AI on board the aircraft, that's primarily meant for navigation.

Intelligent Intelligent navigation for swarming, for detection, recognition, and targeting. And then you can think of AI on the ground station side being more dedicated around the human machine interface using maybe large language models to be able to talk to your drone and tell it what to do without having to manually use your joysticks to be able to do that.

I think AI with respect to drones and in the application of drones just has so much opportunity, and there's going to be so much value to be created here over the next decade. I think there's huge amounts of opportunity for a lot of folks.

- [Ryan] Computer vision obviously is going to play a big role here, right?

Is the ability for the camera to do different elements of this. Where are some of the engineering challenges that come along with the development of these drones for a lot of the applications that we've talked about with incorporating other different technologies like AI tools and solutions as well as computer vision and different things like that into them?

Where have you seen the biggest challenges or what are the biggest challenges we currently are facing right now in order to enable these drones to do certain things that we may be talking about today?

- [George] I'll start off with what challenges every market has in common and then maybe a few specific ones that each one has.

The common challenges across making a small intelligent drone mostly comes in your SWAP constraints. SWAP stands for size, weight, and power. You have a very small drone and to really have the best navigation and computer vision capability, you want this AI to be running at the edge on board the aircraft.

So you need a powerful processor to run these neural networks and models but it can't be too big, it can't consume too much power. You have to really play a delicate balance in the architecture and the design of the drone to be able to still make it small with long range and long flight time, but still have the compute capability on board.

So that's actually a non trivial task and takes a lot of years and a lot of money to perfect. So that's a common challenge across all use cases. And then I think that just gets magnified as you start looking at specific segments like defense, where you want like all of your AI to be running on board the aircraft so that even in these contested environments where drones are being jammed with jammers for radios and GPS, you still want the drone to be able to do what it has to do.

And that just takes more AI and requires more compute power and is generally harder to engineer.

- [Neil] George, what you were talking about made me think of, this is pre pandemic, but the kind of the world of IoT conference, the conference, and I remember in 2019, we had all these kinds of drones and talking about the SWAP, a lot of people were saying the future of drones were like these actually giant drones that have the ability to carry people from location to location. It sounds like that is more of a distant type of we'll say goal or use case than some other possibilities like maybe package delivery.

- [George] To better clarify the spectrum of drones that are out there, what we're talking about right now are Group 1 drones, which are typically pretty small and there's essentially five groups of drones, 1 being the smallest and 5 being the really big ones that can carry really heavy payloads.

I do believe there's a big future ahead of us for eVTOL and unmanned systems that can carry people. Again, I think it's just requiring some regulatory and technology inflection points that haven't quite hit yet, but they are on the horizon. Package delivery is another big use case. Amazon's been trying to do it for a decade. But it is finally making progress, and I also believe it's just around the corner. And then with Teal, we're not focused on either of those use cases, but we're seeing so much on the horizon as well in terms of what even just these small drones are going to be able to do in the next couple of years.

- [Nikolai] I thought it was interesting, you brought up swarms. I don't know if you've seen the film, I think it was endorsed by the computer scientist, Stuart Russell, about killer drones. Essentially, these very small drones that would operate in swarms and would carry an explosive on them and then just fly into a target, like a human target.

Do you see that as being a real possibility? The way the film depicts it is just literally send a swarm to the Capitol Building or something and target senators, right? That would be the threat that if a another nation were to deploy because the drones again are pretty cheap. Is that a plausible future?

- [George] It is. And it's pretty scary. I think there's a lot of things to compartmentalize there and talk about. I think one is there's a great point in there, which is it's really hard to counter this technology. If you have a swarm of drones coming at you, and they're weaponized, if you're on the battlefield, you can't shoot them out of the sky with your rifle.

You can't keep using 10 million dollar missiles to shoot them out of the sky. These are cheap, small, scalable systems that have a really big impact, and this is why people think this is a big future of, a big part of the future of war. And of course, that technology can be used even outside of war, which is scary.

I think there's a lot of ethics to talk about especially with respect to AI. We believe that we should not ban AI or restrict AI from being used on drones. We think it should be done responsibly though and in the context of defense in a way where the operator is always in the loop and making the final decision.

If we look at the American industrial base right now, I mentioned earlier that DJI out of China monopolized the industry basically over the last eight years, crushed the competition. And in the U.S. where we stand today is there's only two drone companies left in this category, Teal being one of them, that is able to scale and manufacture and design this kind of drone.

And we are currently the only ones that are willing to weaponize the technology and deploy it to our service members. So, I think we have a pretty good pulse on what things are looking like and where they're headed. There is always a scary future with any technology, but I think we have a responsibility to be really ethical about it and be careful on how we deploy it.

- [Ryan] Who's responsible for kind of setting the ethical guidelines when it comes to the use of drones and development of drones and in situations where people's lives are at risk or you might be taking someone's life through it as well as I guess the other side is who's responsible if something goes wrong with a drone, right?

Is it the person operating the drone? Is it the manufacturer of the drone? Is it, is who, where does that kind of responsibility fall onto?

- [George] This is definitely where the rabbit hole begins. There are a bunch of good guidelines and principles that the DoD has put together and released in terms of AI and weaponization.

The gist of it is that AI extends the toolkit for a warfighter, and they are still responsible to use that toolkit in an ethical manner. If you think of existing technologies, a Hellfire missile that's being shot off of an aircraft, a pilot isn't manually steering that missile, right?

That missile is smart, and it's following a laser or some other high level command. I don't think we would want to make pilots manually fly every missile they shoot. So this is essentially an extension of that. I don't think we're going to be seeing killer drones that are making all these decisions on their own.

Humans will always be in the loop with their finger on the red button to make something happen.

- [Neil] I just want to reiterate, drones, like all other technology, are a tool. They're not inherently evil or good. It's all about how we as people choose to use them. Is that, am I, is that still true, George?

- [George] Yes, 100%.

- [Ryan] It's a good way to put it. It's always something to bring back. I think anytime you see, what, Netflix has a show, right, the Unknown series, they did a whole thing on killer robots and the use of robots, AI, drones in warfare, and I haven't watched it, but I, so I don't know what kind of angle they're coming at it from.

But when you see things out there like that, it always starts to raise certain people's suspicions on what, to what extent are they going to be used and to what extent are they going to be controllable? Neil, the way you're making people remember that it is a tool and how we use it will determine kind of its value and what it does.

So I think it's very important here, but I think people just go down the rabbit hole of just thinking of all the things that could possibly go bad and go wrong when there's lots of different kinds of content being put out about the potential, whether it's realistic or not, of what could happen.

- [Neil] Yeah, I agree with that, but I feel like I got to make a point. So I apologize everybody taking us on a little tangent. This conversation has come up a lot with various kinds of technology, at least for me, the last two weeks, in that there's a very different perspective of technology, like AI and drones, in Western culture versus Eastern culture.

In Western culture, we're always thinking about the threats, the misuse, right? Nothing wrong with that, but that's our mindset about what's going to go wrong, and it's really the stuff we see in movies and books and stuff like that. But in Eastern culture, they have more of the mindset about all these things are like helpers and tools and assistants that we can leverage to solve problems and help people.

And so they come in with more of a wired mindset of the opportunity to leverage these different technologies. One, I find that a little fascinating, but two, you'll be looking at Eastern culture movies and books and stuff, that's the attitude they adopt. Is that a challenge for us? Perception is reality, as they say.

If we got movies like swarms swarming, weaponizing of drones, swarming, even the Spider-Man movie didn't really paint the things in the best light. Does that limit how we think about the use of some of those technologies, like how we think about how drones could actually be used, whether for military or commercial purposes?

- [Ryan] Yeah, another thing to think about too is, this is, goes back to obviously a while up until now is I've always been a fan of the Star Wars series and that's where I first heard the term drones growing up, right? And you see them used in a utility for certain purposes, but they have their limitations.

And then as the series, as the movies rolled out and as different series were built, those drones started to take on different kind of I would say personalities, but different, they were able to do more and take on a life of their own at times, which kind of created a lot of interesting kind of discussions to be had around where drones are going, which kind of influences what you're talking about, Neil.

- [George] Yeah, these are great points. I certainly think Hollywood has had a big hand in how AI and technology like drones is perceived. But I think every technology goes through this kind of a transition. We're focused on making the most positive impact we can for defense, protecting our country, in enterprise, accelerating the adoption of drones and solving problems that weren't able to be solved before.

And over time, I think people will come around to this as just being another ubiquitous technology that has its own place as a tool.

- [Nikolai] My first exposure to drones was really through like travel videos, the consumer side, and you're right, DJI is, totally dominates all of that.

And they do influencer marketing as well for these travel influencers. But yeah, I've used DJI drones, and they're really great products. They really, they're impressive what they do. Like they're very robust and reliable. So on that note, I guess maybe we could dive into a little more detail about like because right now the U.S. and China, not just around drones, but U.S. and China are feuding about AI chips and there's export restrictions to China and the U.S. military I believe even banned DJI drones because of all the data that the drones collect. So what do you make of the U.S. and China right now and how big of a sort of role is are drones going to play in who comes out on top in the future?

- [George] I'm really glad you asked that. And this plays into what I mentioned at the beginning of the podcast, which is this has been a really unique industry over the last eight years, and primarily because of these dynamics between China and the U.S. So, DJI was the first company to release a ready to fly drone at a really low cost, very quickly spread like wildfire, took over the market.

There were multiple U.S. companies that tried to compete but drones, to be successful and be profitable, take a lot of time and money to do so and China just beat us in that respect for a lot of different reasons. And so they just crushed a lot of the American competition that was around back then.

But what the military then saw and what the government identified is that there's three major risks in deploying DJI drones in operations, both domestically and abroad. There's the data flow aspect, which you mentioned, which is DJI is a Chinese company, they can send data back to China with their drones, which is bad. There's a big dependency issue, which is should conflict ever arise and the only drone capability we have is Chinese drones like, we don't have a capability of our own to meet our needs. And then there's the dual use risk, which is by continuing to procure DJI drones, we're continuing to fund their company, and thereby China's drone industrial base.

So we believe, and the government believed here in the US, that we need to rebuild the American drone industrial base, counter and mitigate those risks and have a drone that, a drone capability that ultimately will give America the asymmetric advantage and deter conflict before it ever happens.

- [Nikolai] Yeah, I think some people don't realize that like every company in China is required by law to basically report to the CCP. So yeah, those are, everything you mentioned, those are all like, the connection back to Shenzhen is going to be a risk for the U.S. military.

- [George] Exactly. Yeah, no, it very much is and what the U.S. Army did is in 2017, they were the first branch to ban DJI drones. They very quickly put together a program of record called Short Range Reconnaissance, which is made to give a drone to every platoon in the Army and that's the program that Teal joined back then and what spurred us to pivot to defense. And so now that we're seeing and that the world is seeing the impact of drones, we're seeing a lot of capital being allocated to this technology, and I think this decade is going to be a massive decade for growth.

- [Ryan] Speaking of that, I wanted to ask you one of the last questions before we wrap up here, is just what does the future of drones look like? Like where are we going? We've talked a lot about where we started, how we got to where we are now, answered a lot of really interesting questions that I think a lot of people out there are probably curious about, but where do you see drones going forward, is, in both sides, on both the consumer side as well as the government military side of things as we look out three, five years into the future or whatever horizon makes the most sense to be thinking about.

- [George] I'm very optimistic. I think there's a very bright future for drones, both in terms of what they will be able to do and how they'll be made. I think swarming is a big aspect to this again, being enabled by AI, having one person fly many drones at the same time is a massive inflection point.

I think another one is if you imagine a drone that can intelligently navigate an environment, not crash, be able to be piloted from the other side of the country or remotely from anywhere and then be able to take off and land from a basically robotic box that can store the drone and is connected to the internet and can deploy the drone from anywhere at any time.

You can imagine just a whole suite of use cases and applications. They can take advantage of that. So I think the future is undoubtedly autonomy, AI, and there's a lot of these regulatory aspects to it that I think will unlock a really bright future.

- [Ryan] Are there any kind of, just quick, use cases that you feel drones are going to start to play more of a role in the enterprise setting?

So not necessarily for like community and natural resources and kind of the environmental stuff or the defense but just more generally for certain industries just on a more regular basis.

- [George] It's hard to think of an industry that won't be able to be positively impacted by drones. If you think of them being able to be autonomous and powered by AI being flown remotely, you can think of them being used for inspections for a bunch of different types of infrastructure. Neil mentioned public safety, police, fire, search and rescue, wildlife conservation, so many different things that with all these different technologies coming together into a drone, I think enables some really interesting things.

- [Ryan] Neil, any last questions or thoughts from your end?

- [Neil] I think it's actually been a great insight to learn I think a lot about different types of drones, how there's a convergence with drones, AI, and actually IoT.

But, I know a lot of people naturally gravitate to drones and military thinking about like weaponization, but it's great to actually hear some of the other use cases about helping protect our soldiers lives and other uses, especially for public safety like wildfires.

- [Ryan] Yeah, very exciting times, very optimistic about what this is going to lead to.

And we really appreciate your time, George. This has been a great conversation on a topic we haven't had a chance to talk about yet. For our audience who wants to learn more about what you're all doing at Teal, as well as just maybe follow up with any questions or thoughts, what's the best way that they can do that?

- [George] Feel free to check out our website at tealdrones.com and reach out to us through there.

- [Ryan] Really appreciate your time and excited to get this out to our audience.

- [George] Thanks everyone. Really appreciate it.