Brandon Tseng: “I don’t think there was such a transformative technology since there was the development of nuclear bombs.“
Kalenderwoche 23 // US-Perspektive
Brandon Tseng served as a SEAL in the US Navy, commanding several missions in Afghanistan. Back home, he starts developing the idea of autonomous drones, piloted by artificial intelligence. His vision is to apply masses of drones on the battlefield. Nine years later, his company raises 800 million US-Dollar funding, the company is valued at 2.7 billion US-Dollar. In his first interview with German media, Tseng explains his perspective on drone warfare, the comeback of mass on the battlefield and why he sees in drones and artificial intelligence the largest game changer since the development of the nuclear bomb.
To begin, please explain us your view on the current situation of global conflicts from the two perspectives you bring together, as an entrepreneur and a former US Navy Seal. What meta trends do you see in conflicts?
Thank you! Overall, we see a shift from special operations to land taking operations at a larger scale. I would like to talk first of all about a specific thing, that highlights the bigger problem very well: The US or its NATO allies are sending F-16 Fighter Jets over to Ukraine. However, I think this is more of a political statement and a moral support because they cannot really fly right over the front line. There are so many air defence missiles right at the front line. I have been talking to the F-16 program manager in Ukraine and the pilots are stating all the same: They would not have any chance in direct troop support. Electronic warfare and GPS jamming have been rising so much. At the same time, Ukraine is losing about 10.000 drones per month. I have been talking to the national security advisor, and he tells you: We need to figure out, how to operate without communications and GPS.
Legacy platforms are becoming battlefield irrelevant. That’s what I am telling everyone in every nation. I would not buy an MQ-9, an MQ-4 or a Reaper drone anymore, because they are very expensive platforms and they are very cheap to eliminate now, simply by mass. At the end, we would not use these items in a hot conflict anymore, neither in Ukraine nor in a hot conflict against China. We can call this a meta trend, air superiority as we know it is getting under pressure.
So why did the United States initially buy these systems, when they cannot use them in a hot conflict?
There has been a transformation. Prior to 9/11, we have been fighting with fighter jets and helicopters. Then from 2001 to 2020, we learned how to fight with Reaper and Predator drones, which were extremely effective against al-Qaida, ISIS and global terrorist networks, which were not able to apply sufficient air defence and we established air superiority. But in scenarios where we have no air superiority, these assets are useless. You can look it up, we lost nine or ten of our jets in the past 16 months. I have spoken with senior authorities in the US Air Force, and they say: We are sick and tired of losing a 40 Million Dollar aircraft to 500.000 Dollar missile. At the end, war is about treasure, and this really hurts the treasure. So, coming back to the meta trends. There are a couple of factors, but one meta factor is, that we are losing air superiority. Why is this so important? It allows you to move your troops on the ground. The last war, where we had no air superiority was World War one. A terrible war, in which it was incredibly hard for each side to make progress. You are seeing that in Ukraine as well.
This also leads to political and economic meta trends. When I was serving in Afghanistan ten years ago, I saw that the world was more stable than now. China is claiming Taiwan, Putin obviously is attacking Ukraine, Iran is backing Hamas and the Houthi Rebels in Yemen, which are not only attacking US and allied ships, but also the international trade routes.
There are now too many missiles and not enough aircraft in the world, in combination with electronic warfare. When we had our eye off the ball to fight the terror, Russia and China were looking at us and saying: Ok, what does the US rely on? It's air superiority that consists of fighter jets, helicopters, drones. Their solution was to invest in constructing missiles, jamming GPS and building up electronic warfare. This is why we as Shield AI are focussing on AI pilots and low-cost aircraft.
Ok, let us assume there will be no human pilots in the near future? Coming back to the meta trend and the cost factor of unmanned systems. Do economics rule future conflicts more than in the past?
War has always been about blood and treasure. So economics have always ruled conflict; but in the future it is more about production capacity and regeneration rates (how quickly you can build drones) so the economic factor is even more important. Smaller military budgets will go much further in the future because the future will be about how smart your drones are and how fast you can build them. As I said, at Shield AI, we are focussing on AI and low-cost aircraft so solve this problem. You have the P-8 , that’s a 180 Mio. Dollar aircraft; you have the MQ-4 Triton, that’s a 180 Mio. Dollar drone; you have an MH-60, that’s a 40 Mio. Dollar helicopter. The helicopter and the P-8 require air crews. The same thing is happening with fighter jets in the world. We have the F-35 and F-22, but we are moving to a world, that we call CCA (collaborative combat aircraft), the Europeans call it ACP. There you have cheaper collaborative platforms. The XQ-58 by Kratos for example, is a 6.5 Mio. Dollar Aircraft, while the Platforms I mentioned before, are costing more than a hundred Mio. Dollar per aircraft. However, you can move to a world where you can have many, many more cheap drones, piloted by AI. AI does not require GPS, does not require communication. So, you can go at low cost into these conflicts and especially, you do not have to worry about human lives if they get lost.
Would you say, the development of a complex system like FCAS makes even sense anymore?
Frankly, as I understand FCAS, it has a crewed and an uncrewed variant, as the F-35 Program in the US. The crewed variant makes no sense. In the future, I see a few unmanned F-35, surrounded by hundreds, if not thousands of drones. I think some European leaders also are starting to understand that. In my opinion it makes zero sense that we think about a crewed variant. My 3-year-old and 1-year-old will never have to drive a car because of self-driving technology. Should my kids choose to pursue the same path that I did and attend the US Naval Academy, I don't believe becoming a fighter pilot will be an option for them like it was for my generation.
Can you sum up this as what you see as a meta trend?
Sure, that’s what I see as the meta trend: You have very expensive drones against very cheap weapon systems. A Chinese Hóng Qí-9 missile is probably 1 Mio. Dollars, going against a Triton, a 180 Mio. Dollar drone. Every time they get shot down, you can say: Great, that’s a 180. Mio. Dollar to China, minus 1 Mio. Dollar to the US. On a meta base, you have to flip that cost equation. That is a massive point in great power competition. You will need cost advantage to sustain any future conflict, and we are not even near that. Especially as China as about ten times economic advantage over the US and still three times economic advantage over the United States and the West combined. Only by closing that gap, you can prevent a conflict. The Russians, China, the Iranians have to know, it is just too expensive to go to war with us.
What is the solution to this, in your opinion?
You want to move to a world with hybrid warfare. I am not saying you are getting rid of every single pilot in the next ten or fifteen years. But you will have far less expensive assets. Up until 2001, there was the domination of fighters; then there were expensive and complex drones; now we go to cost-effectiveness, or intelligent, affordable mass.
That’s a hypothesis, can you explain that thesis further?
Oh, yes (laughing), it is more or less proven all the time in the recent history of technology. Let’s just take computers: We started off with huge desktop computers with lot s of human interaction. Then we went to laptop computers, now everybody has mobile phones that do the same job. So, desktop computers have been the fighter jets, Predator and Reaper are laptop computers and intelligent drones are our mobile phones.
This example automatically leads me to think about technological dominance. In exactly the field you are talking about, we started the invention, but now we are losing the dominance. Do you think, this is also happening in AI and UAV technology?
There is a great danger at least. In countries like India, Africa and so on, they skipped desktop computers and laptops! Everything is built off the mobile phone. And now they are straight building up affordable intelligent mass. Larger and established militaries, like the US and Great Britain are having a harder time to jump onto that concept of affordable mass. Now you must figure out how to make that transition, while you have the existing forces. It is like when a company cannot change quickly because they have so many existing capabilities. Whereas a country that has a smaller budget… let's say India, Brazil, they haven’t yet spent that much attention on these very expensive platforms as we did. I am confident, in five to ten years, you will be able to give a platoon the power of a fully equipped strike group (Anm. d. Red.: Flugzeugträgergruppe). This will change the world.
You are talking about the chance that the US is losing military world domination in five to ten years, do I understand that right?
Possibly. I don’t think there was such a transformative technology since there was the development of nuclear bombs.
That’s a big comparison. How can drones be as disruptive as the one nuclear bomb?
Well, what was the situation then? When the Germans discovered nuclear fission in 1938 that ignited an atomic race and the Manhattan Project. At that time, the US realized very quickly: We need to win this race, no matter what. The one who wins this technology, will have such a significant advantage, because this outruns all common technologies of that time. At the end, the outcome of this race set the world order for 80 years! And I can see something happen like this with AI pilots, drones, and affordable mass, too. The one who has mass and cost advantage on its side and does not even exposure its own people, will probably set the world order for the upcoming 80 years. It is how I think about the world, from my understanding of war on a meta level.
Do you think that China invested in AI and UAVs quite early, because they had the foresight to think in the concept of mass and cost advantage?
I don’t know if they had the foresight to do this. However, what China has, is an industrial base like no other. Mass is a fundamental concept of war since there is war. Who has more troops and more weapons, tends to win more battles, tends to win more wars. China demonstrated that they understand this very well in the Korean War. They sent a million people over the Yalu River and there were simply not enough bullets. The Chinese understood that from history and I think they never let that concept down, while the West tried more and more to win battles with less mass. To win special operations like we did with the Seal Teams, ok. That works. Our missions were very specific. But actually, seizing a landscape with 16 people, that does not work. Afghanistan is the size of North Carolina. Yet we could not seize it, even though we won the special operations all the time. So, the West thought, they could get away with not having mass. Now the idea of seizing countries comes back, that is changing. The Russians also understand the concept of mass very well and that is, why they are making progress in Ukraine. They simply have a mass of army and a mass of material and even worse, they are backed by China and Iran in terms of weapon systems. That’s why it's such a dangerous situation now and that’s, why the Russians are so hard to stop. The same applies for the Chinese.
How would you suggest countering this?
Well, that’s a complex task. I had to battle track 150 Afghans, 16 Navy SEALs, hundreds of civilians, five aircraft in a stack, and the enemy. Keeping track of the red force, blue force, and neutral force picture; and intelligently manoeuvring the mass of troops on a highly dynamic battlefield is a hard thing to do. AI pilots commanding masses of drones will be incredibly effective and efficient. The systems can be trained to command masses of drones. If you use AI to command-and-control masses of drones, you are not limited by personnel anymore. You could have 16 people, but they control 10.000 drones. That’s the concept of mass applied. The only limit then is how fast your industrial base can pump out these drones.
That sounds like a great pitch for future AI and drones, but where are we actually now, when it comes to these technologies?
That’s the point, it is already there. I am not talking about any random vision or idea that we one day want to build! In 2022, we flew F-16 fighter jets completely autonomous – successfully! We just flew the secretary of the US Air Force autonomously! This is no longer theoretical, all what I have been pointing out before, can be done. However, your question tells, how this truth is not really understood by now. My concern on the other hand is that China understands this quicker than we do. If they did, we have a problem. They are good at the mobilization of resources at scale. The west has its bureaucracies and systems in place, which makes change more challenging.
What would have to happen in your opinion now?
The amount of money that we have allocated by now in AI is probably 1/1000 of what it would have to be. At the moment, we kind of try to start a Manhattan Project with 30 Mio. Dollar. 30 Billion dollars was the actual spending of the Manhattan Project, considering the buying power of today. In short, we need to put more resources in the field of AI.
Coming from Germany: Do you think Europeans understand the problem even less than the US?
I guess also the Europeans understand the problem. However, coming to a solution, is way harder. Most definitely, it is not the solution, to burn some 10.000 Drones in the Ukraine month by month. A problem is the allocation of resources and there, I think the Europeans and the US still work in similar ways. You can give 10 Billion dollars to Lockheed Martin and they will build a great drone. But not as quick and cheap in the cost per drone as we need them.
What if technology does work 99 percent, but not 100 percent? If you have just a little bug, wouldn’t then regular forces be more reliable?
A lot of people share that concern. And don’t get me wrong, we need to change quickly, but not 100 percent at one time. Compare it to World War II and the navy. Battleships were extremely powerful. However, with naval aviation coming up, you did not want to be on a battleship, when 150 Hellcats came over the horizon. This led to new kinds of ships and carriers. A challenge in the next years will be definitely the change in the bigger forces, like the US, UK and Germany. My point is, they need to identify the objective, to step by step becoming an AI-piloted army, navy, and air force. Not with one day of change, but a fast and steady transformation.
Let's look at this transformation from a political view. The more data you have, the better AI becomes. This would indicate to put all western strength together. On the other hand, there is national sovereignty and under Donald Trump, the relations between Europe and the US were less stable than before. He might become the next president. What would be your advice? Putting all industry forces together, or stay in competition?
First and foremost, I think alliances are critically important. Whether that’s NATO and Europe, Australia, the Philippines, Japan, and Korea with the United States, it is critically important. Everybody has to understand that. The second thing to consider is, only a few companies in the world are able to build a fighter jet. Boeing, Lockheed Martin, Northrop Grumman, and Airbus. However, many, many more companies can build cheap drones. We are seeing that in Europe, we are seeing that in Ukraine now. When it comes to the software expertise, I really think the United States are the best in the world to build software and AI. I would not really invest much in other nations when it comes to software. What I would lean on, are in fact alliances. It is just starting, but the US, UK and Australia are looking at licence free defence export. Basically, this could be a trade free zone between these states, and I would expand that. Everyone can contribute to that market, in the building of the drones. The software is very hard to replicate, that’s why prefer the idea of alliances. Also, this sends a signal to Russia, saying “we are unified”.
Coming to the ethical question. In the past, you had a tank crew that was trained by the state. Hence, the state guaranteed that he trained the personal using the tank firepower right. If now soldiers operate with 10.000 AI piloted drones and some of them attack non-military targets, the state might say: Hey, the company guaranteed to us, that this software is working perfectly correct! Talk to them, this is not our fault. How would you deal with this?
First, as a US Navy Seal, I am very well aware of all the ethical issues of using lethal force against the enemy. The decision is always a very human decision. And I would absolutely say that the decision to use lethal force always has to stay in human hands. It is the moral high ground, it is the ethical high ground, I think this is super important. Some people might say, it is a tactical disadvantage to actually make this decision, because warfare is becoming faster and faster. It doesn’t matter to me. I think it is a strategic advantage to have the moral and ethical high ground. It is simply a question of who we want to be and what our values are. And as far as I can see, our allies, the NATO think the same way. Even Xi Jinping has said that he thinks, AI should not make killing decisions. Well, maybe take what he says with a grain of salt. I don’t know. The AI systems are not designed to make these decisions themselves. That’s also why I think, an F-35 is still important on the battlefield. You need to have a human soldier with a good picture to make the decisions. That’s how I think about it.
In combination with this answer and your comparison of AI to nuclear weapons, do you think the world is in need for an international AI agreement how to use AI and do you think there is a realistic chance to achieve that?
(Thinking.) That’s a great question. I don’t think you will see that happen. There is a big difference: Nuclear weapons are widely destructive if things get out of hand. One decision, and you can destroy an entire city in some seconds. Every country who that has nuclear weapons recognises it. That’s why there are a lot of checks behind it. The difference to put up a million drones on a battlefield is, that it is incredibly controlled. The comparison here is more like you would say: “AI is not allowed on the battlefield because you are not allowed to use a million people on the battlefield.” That’s why I think in this aspect there is a difference between AI and the nuclear bomb. And I don’t think you will see any countries not using it. In fact, it will make war in tendency less destructive. The nuclear bomb made it more destructive.
Thank you for this interview and your thoughts, Mr. Tseng.
Das Interview führte für Lagebild Sicherheit Dr. Christian Hübenthal
In der wöchentlichen Ausgabe erhalten Sie kostenlos das vollständige Lagebild der Sicherheit.