Are We Ready? | The New Weapons of War, With Michael Horowitz

LINDSAY:
The world has turned dangerous. Is the United States prepared to meet the new challenges it might face? In this special series from The President's Inbox, we're bringing you conversations with Washington insiders to assess whether the United States is ready for a new, more dangerous world. Technology is changing the face of warfare. Cheap drones, robotics, sophisticated sensors, and artificial intelligence are among the advances that have militaries around the world scrambling to rethink their force structures and doctrines. Whoever innovates the fastest may gain a decided advantage on the battlefield.

https://www.youtube.com/watch?v=H_GfnFzhd4w
NBC News:
In the future of warfare, Europe is rethinking how to defend itself. The continent is scaling fast with smaller systems, smarter tech, and even some private money footing that bill.

https://www.youtube.com/watch?v=asEMh_6BXsU

Bloomberg Television:
It's really about a changing nature of warfare where we're looking at how to incorporate autonomy into all kinds of different operations.

LINDSAY:
For more than eight decades, the United States has operated at the cutting edge of technology. Today, however, entrenched bureaucracies, outdated procurement systems, and an attachment to legacy systems may be slowing the Pentagon's efforts to incorporate new technologies. Whether the U.S. military adapts to change more effectively than its rivals may be the difference between victory and defeat in a new age of intensified geopolitical conflict. From the Council on Foreign Relations, welcome to The President's Inbox. I'm Jim Lindsay. Joining me today is Michael Horowitz, the Richard Perry Professor of Political Science at the University of Pennsylvania, and a senior fellow for technology innovation at the Council. Mike, thanks for joining me.

HOROWITZ:
Thanks so much for having me. Looking forward to the conversation.

LINDSAY:
Now, Michael, we decided we want to do a piece on technology. I really wanted to talk to you given your background, both as a scholar, but also your work in government. You recently served as Deputy Assistant Secretary of Defense for Force Development and Emerging Capabilities, as well as Director of the Emerging Capabilities Policy Office. So let me ask you the big picture question to start, how significant are these changes that we're seeing in military technology?

HOROWITZ:
I think the changes that we're seeing in military technology that have been brutally demonstrated on the battlefield in Ukraine, that we see in the Red Sea, that we see in the conflict in Israel and Iran, definitively show that a change in the character of warfare has occurred. I think we have entered the era of what I've called a precise mass in war. And by that, what I mean is the combination of artificial intelligence and autonomy, the diffusion of precision guidance capabilities, which are now fifty years old, essentially, and commercial manufacturing, mean that now essentially any country around the world, and sometimes sophisticated militant groups like the Houthis, have the ability to generate precision strike of one sort or another.
And it's not that every missile that every country has, or every one-way attack drone, is maybe as good as a Tomahawk missile in the United States. It's not. But it is a fundamental change that so many different countries now have access to even simple versions of these capabilities. And what we've seen in Ukraine and Russia, in the Ukraine-Russia context, is that these capabilities are really changing the way that militaries fight both at short ranges and at long ranges. And so this is exactly the kind of disruptive situation that can transform the military balance of power.

LINDSAY:
So Michael, help me understand this in historical perspective, because revolutions in military technology are not new. Actually, over the weekend I was reading a book focused on the run-up to World War I, and it went on at great length about how much technology had changed. You had the invention of motorized vehicles, tanks, and aircraft, advancements in explosives, radio communications, chemical weapons, and of course the rise of a new classic battleships that really shook up particularly the British and the British Navy. World War II, we obviously came out of it with nuclear weapons. You mentioned precision-guided munitions. I remember back in the '90s with the Gulf War, that precision-guided munitions had revolutionized the way wars would be fought in the future. So how different are things today?

HOROWITZ:
I think things are increasingly different, but just to be clear, it doesn't mean that everything has changed. I would say, I don't love using the word revolutionary for... If you grant me a brief tangent into history, in that in the 1990s, you mentioned when precision guidance first came of age in the public eye, you had all these conversations about whether the information age was yielding what was then called a, quote, "revolution in military affairs" for the United States. And the challenge was that whole conversation ended up getting buried under the weight of how to define the word revolutionary. And so in some ways, I think the question of whether it's revolutionary is less important than what are we actually seeing change and how? And what we are seeing change is actually quite substantial.
And here's an example. For the last forty or fifty years, militaries like the United States that wanted to strike targets, because of precision guidance we shifted from a world where in World War II, you'd have bombers fly over a city and a dozen bombers might drop dozens of bombs each to knock out a single factory, to a world where maybe one bomber would drop some bombs, to a world where one plane would fire one weapon, to a world where if a country like the United States fires a weapon, the idea is, "All right, on the third story of this building on the left side, we want to hit the second window from the left because that will destroy the part of the building that we wish to destroy and nothing else." That way of thinking about how to fight made perfect sense in the context of maybe something like the war on terrorism or the 1990s. But in a world of great power competition or even just in a world of conventional warfare, you'd quickly run out of weapons.
And indeed, we've seen the United States military get smaller, more expensive, and with the unit cost of all of those systems going up. And so that's a military that might not be quite as well-suited to the kinds of great power conflicts that many people believe could be coming on the horizon. And simultaneously, other countries, because as I said before, now have access to all of these precise mass capabilities, they can then use those to impose costs in ways that militaries could only dream of before. Here's an example. In the Red Sea for almost two years, the Houthis fired what could really only be described as flying lawnmowers, sort of $20,000 systems with precision guidance, at U.S. ships that, frankly, unless they were shot down, would hit their target a lot of the time. And against those $20,000, maybe $50,000 flying lawnmowers, the U.S. Navy was shooting $2 million missiles to stop them.
That is a cost exchange ratio that can work well, I guess, when you're a really wealthy country like the United States. But in the context of, if you have imagined a war between the United States and China, a war between the United States and Russia, or even frankly, the interaction between Russia and Ukraine, that kind of cost exchange ratio doesn't work anymore. And we now then increasingly see not just that, but one way attack drones, millions of them being produced by Russia and Ukraine to launch at each other and for surveillance in ways that just lead to a very different battlefield than the one that we've become used to over the last forty years.

LINDSAY:
So Mike, is the argument that the United States military is going from a position which it had a qualitative advantage over others, to a situation which its qualitative advantages can be overwhelmed by the quantitative advantages of lower cost technologies?

HOROWITZ:
That's a great question. I think there are two things going on. The first is there is no country in my lifetime that has produced a military challenge to the United States as severe as the one that China poses today. The sophistication of China's military technologies, the way that China essentially has hit full send on all military capabilities at once and can produce them, means that there's more pressure on the United States at a high end than ever before. Simultaneously, because the United States has been so focused on having small numbers of exquisite capabilities, it's not just that the qualitative edge is eroding a little bit with regard to China, although there are areas obviously where the United States is still way ahead, like say, under sea warfare, but it's also possible then to overwhelm defenses in just the way that you said. Using, say, precise mass capabilities to overwhelm air defenses or overwhelm ship defenses. This is something we see Russia do to Ukraine on a weekly basis, frankly, in which imagine what that would be like in the context of a future great power conflict.

LINDSAY:
Well, no, Mike, I was just going to ask before we leave this topic on the issue of history, are there any lessons from previous, I won't call them revolutions, but let's say surges in technological capability that we should keep in mind as we think about the situation we're in today?

HOROWITZ:
I think the biggest lesson we see from history, maybe the two biggest lessons that we see from history macro, one is that history is littered with leading military powers that fall behind because they rest on their laurels, because they look in the mirror every day and say that they're the best until they're not. We see this with the British and their battle cruisers getting sunk at the outset of World War II by Japanese naval aviation around Singapore. We see this with the way that English longbows defeat the mounted knight. We see this again and again in military history. And so these kinds of periods can be really risky, frankly, for the leading military powers if for bureaucratic reasons they're too slow to transform to the new security environments. That's thing one.
I think thing two, is that the winners in these periods tend to be militaries that are not just wealthy, but have high degrees of what I would call organizational capital. And by that, I mean the ability to not just pivot on a dime, it's very difficult for any large military organization to do that effectively, but to make change in a timely manner that's consistent with the degree of change that we see in technology. Because victory and defeat in warfare is not just about having the best technology. Again, we see that again and again on the battlefield. It's not just about having more. It's not just about having the best. It's about who can figure out how to use capabilities the best. It's that intersection, basically, of technological and organizational change. And so you can't just be big but not good, or good and not big, especially in an era where everybody will be able to mass produce some element of precision strike.

LINDSAY:
So Mike, is there an advantage to being the first mover in adopting new technologies, or is that sort of a red herring that we shouldn't be worried about?

HOROWITZ:
I think first movers can often get advantages when it comes to the introduction of new military capabilities, and those advantages are exaggerated when you have really high unit costs, when the cost of each individual thing that you're buying is really high, or when they're really complicated military technologies that, frankly, only militaries would build. One example of this is a stealth technology. Stealth technology is, both the United States and the Soviets are working on it. The United States makes the technical breakthroughs to get there, much to the chagrin of the Soviet Union. And it provides an enormous first mover advantage for the United States because stealth technology is really expensive, really complicated, and there is no commercial market for it, which means that militaries have to do the research and development and procurement, production, etc, themselves. You're not going to have lots of private capital investing in it to try to succeed.
All that is to say, it's not that you always have to be the first mover to take advantage of different kinds of capabilities in war. Look at the Prussians, for example, the armies of Bismarck that dominate the French in the Franco-Prussian War in 1870 and reshape the European balance of power prior to World War I. The individual technologies that the Germans were using, the most prominent for the railroad, rifle, telegraph, the Germans invented none of those. What they did was figure out how to use them effectively in these open order tactics that allowed them to overcome French forces and reset the balance of power in Europe. So one, yes, there are first mover advantages. And the more expensive things are, the more exclusively military they are, the higher those advantages are likely to be, which makes me nervous, frankly, about the technologies of today for reasons we could talk about.
Two, you don't have to be the first mover in a technology to take advantage of it from a military perspective. And we have lots of great historical examples of, frankly, leading militaries that are not necessarily the first to invent a technology, or leading militaries that invent a technology but are not the most effective at using it with the British inventing aircraft carriers and then viewing them as a spotter for the battleship in World War II, as a really classic historical example.

LINDSAY:
So Mike, as you surveyed the terrain today, what are the technologies that you worry most about?

HOROWITZ:
I think artificial intelligence is profoundly likely to alter the character of warfare, frankly, in ways that could be even larger than the precise mass era that I've described us as being in now. Since precise mass is not necessarily... It's not about fancy LLMs invented by cutting-edge frontier AI companies.

LINDSAY:
LLMs being large language models.

HOROWITZ:
Oh, sorry. LLMs being large language models or generative AI. You don't need ChatGPT to launch precision weapons, as Ukraine frankly shows. That kind of AI is not necessary. As you have continuing advances in artificial intelligence though, we imagine that these are likely to reshape the economy and society and militaries. And I think of artificial intelligence as a general purpose technology, which is one reason why we were joking before we started here, I have this huge stack of books on late 19th, early 20th Century U.S. defense investments, German, British defense investments, etc. That's the last period of time when you really had, to the point that you were making, all of these general purpose technologies coming online at the same time, the combustion engine, electricity, the airplane, etc.
And those both have a short-term impact on war as countries are figuring out how to employ them essentially from the drop, but also because of the way that they then reshape economies, influence productivity and economic growth, and then have a long-term impact on both the technological capacity of militaries and their economic potential to invest in militaries. And I think that the era that we are entering now is one where artificial intelligence as a general purpose technology is profoundly impacting everything.

LINDSAY:
Can you give me an example-

HOROWITZ:
And so we should see it in that way.

LINDSAY:
... how do you see it changing things? How do you see it changing things, Mike?

HOROWITZ:
I see it changing, I think there's several different layers through which I could imagine artificial intelligence reshaping the way that militaries behave and act. The simplest one, frankly, has to do with logistics and with the use of tools like ChatGPT, the ability to use artificial intelligence to improve logistics, to improve the ability of militaries to design supply chains, to do acquisition paperwork. There are real opportunities for returns on scale in the bureaucratic back office of how militaries operate. Then there are potential advantages in what you'd call decision support. The idea in a complicated battlefield environment where you have lots of different assets, lots of different platforms, lots of different weapons, lots of different sensors, trying to put together information, not just faster but more accurately in a way that helps commanders make good decisions in the context of wars. And then there's also risk there if they fall prey to things like automation bias when they trust algorithms more than is warranted, given how good an algorithm is.
Then the third area I think is all the way at the cutting edge. People spend a lot of time worrying about killer robots or those kinds of things at the edge of the battlefield. I think the human's going to be making decisions about warfare for a long time, including, frankly, in an era of autonomous weapon systems, but in a world where you can use algorithms to more accurately target and to determine whether something is not a valid target. Relative to the performance of humans and that increases, the role of the human will change on the battlefield just as it had before. You add all of those things up, and that will then lead to lots of changes in everything from the way that you design the future of military, to the kinds of people you recruit into it and promote, to the kinds of systems that you want to field.

LINDSAY:
So Mike, if AI is the game-changer, is that good news for the United States? I frequently hear that U.S. computer companies are on the cutting-edge of AI. Or is this something where it doesn't matter who invents it first, it's going to be a question of who can figure out how to deploy it in the wisest most sensible way?

HOROWITZ:
That's a great question, and I will say there is a real debate about this. There are definitely those that have argued that breakthroughs in AI, and from the frontier labs, meaning we are headed toward a world of artificial general intelligence, and who gets there first, will be able to potentially lock out others or that kind of thing. I think I am most associated with a different argument that flows from this idea of AI as a general purpose technology. And if you look at some of what I've written for Foreign Affairs over the last year, Radha Plumb and I... Dr. Plumb was the former chief digital and AI officer for DoD, wrote a piece about how it's really about AI adoption. That's what's going to matter for winning. You can have the best models in the world, but if the U.S. government and the American military intelligence community are behind in adoption, it won't matter.
Here's a way to think about it. There's this common saying in military terms that amateurs do strategy and professionals do logistics. The idea that anybody can sit back and be like, "Oh, send the forces that way. Send the forces that way." But are they fed? Do they have supplies? The logistics are what the professionals do. We made an argument in that article that amateurs talk about model parameters, and professionals talk about adoption. So that's the first piece. The second piece, and Lauren Kahn and I wrote an article in Foreign Affairs that was published about a month-and-a-half ago on this-

LINDSAY:
We'll put these articles on the podcast page for The President's Inbox so people can find them.

HOROWITZ:
Great. And that argues that essentially everything we know about the history of technology suggests it is unlikely that there's going to be one single AI breakthrough that reshapes the world and that there's no coming back from. Instead, you're likely to see consistent competition in the AI space because of the huge economic incentives, which are bigger than any incentives that militaries can provide for investment, which means you will have continuing back and forth between AI companies. Which is good, frankly, because it means that if you do fall behind from an adoption perspective, you have the ability to catch up, but it means that AI competition is likely to be a marathon rather than a sprint.

LINDSAY:
So Mike, I want to pick up on a point you mentioned a while ago about autonomous systems, and there's often a debate over whether that is a good thing or a bad thing. And you alluded to, I guess what we would call the man in the loop question or the human at the helm question. You seem to be skeptical that we're going to take people entirely out of decisions about using kinetic force. Are you speaking just about the U.S. military, or are you confident that all militaries or all pure militaries, Chinese, Russian, are going to make a similar sort of calculation?

HOROWITZ:
It's a good question. I mean, from a U.S. perspective, I'm pretty confident. And part of the confusion here is you'll hear lots of senior leaders in the United States talk about, "Oh, there'll always be a human in the loop." What they really mean is a human in the operational loop, because there's a requirement that there always be a human responsible for the use of force. Otherwise, how would you have accountability when things go wrong? That is different than a human always being in the loop at the tactical level to the point of impact. After a soldier fires a gun, we don't require supervision over the bullet before it hits a target.
And I think that there's this myth sometimes because of the ability of the U.S. military and the war on terrorism, to be able to have drones circling above the battlefield and then you could fire a missile, watch it all the way to the point of impact, and there was no chance it would be jammed or disrupted or anything like that, that's unlikely to be how a complicated battlefield would look like at, say, in a U.S.-China conflict where you'd have lots of jamming and electronic warfare. Which in some ways makes more autonomy and weapon systems an inevitability, because you have to do that if you want your weapons to work if they're going to be jammed. But I'm very confident the United States will always have a human responsible for the use of force.
The thing I worry about more is some more autocratic countries, in that autocratic countries do not trust their people to begin with. If they did trust their people, they might have different forms of government. And because of that, I think that there is more of an incentive in more autocratic countries to try to use technologies like artificial intelligence to cut people out. In the United States, the U.S. military talks about AI in the context often about human-machine teaming. How do we use technology to improve the ability of America soldiers to perform and excel on the battlefield? But if you don't trust your people in the first place, there's a lot more of a temptation, I think to cut people out in different ways. It's not to say that accidents can't happen in a U.S. context, but I'm much more worried about some of the negatives that people worry about when it comes to AI in the battlefield from countries that I think have stronger incentives to cut humans out.

LINDSAY:
So Mike, as you survey the terrain, how confident are you that the Defense Department is moving quickly enough, perhaps not too quickly, to adopt new technologies, whether it's AI or any other new technology? And I ask that against the backdrop of, one, we had the passage earlier this year of the so-called Big Beautiful Bill, which allotted, I think it was like $150 billion for new technologies. And I also know that Ash Carter, when he was Secretary of Defense, that was a big fan of basically adapting to new technology. We had the creation of the Defense Innovation Unit. But I've heard mixed reviews as to how successful that has been in changing the way the Department of Defense operates.
Because again, it is a huge bureaucracy, has all kinds of rules for very, very good reasons. This is not a knock on the quality of the people who work there. But as you look at it, are we making really good progress? Making progress fitfully in some areas, not others? What's your grade as a college professor, what you witnessed during your time there?

HOROWITZ:
Maybe like a C+ to B-, and sorry about the slash grade. I think that the-

LINDSAY:
Is this on a sort of state college grading scale or this is an Ivy League grading scale?

HOROWITZ:
It's an A with great inflation. I think there is growing recognition within the United States military at very senior levels of the uniformed military and civilian leadership, of the way that the character of warfare is changing and the need, for example, to invest way more in these precise mass systems than the U.S. military has today. When I was in the Pentagon, for example, a bunch of different offices in the Pentagon worked together to launch something called the Replicator Initiative, which was designed to field thousands of low-cost, attributable, autonomous systems in the Indo-Pacific by August 2025. And that program has had some successes, but there have been criticisms of it, both some of the technologies involved, and whether there was enough money in the program.
And frankly, I advocated for 10x the funding for it. And you know what? The Big Beautiful Bill actually delivered that in some ways. The money in the defense reconciliation part of the Big Beautiful Bill contained, frankly, investments that I would've called my wishlist when I was in the Pentagon in terms of money for testing and evaluation for artificial intelligence, money for scaling low-cost cruise missiles, money for scaling other precise mass kinds of capabilities. There's real value there. There's now a question though, about the ability of the Pentagon to follow through, both given some of the issues between the Pentagon and Congress at present, and obviously we are in the midst of a government shutdown of right now, although certainly that will end eventually.
But what that points to, I think, is there is growing desire and interest in addressing this challenge, both at senior levels of the military and in Congress. But in a world where for forty years most of the budget has gone towards very small numbers of exquisite capabilities, has gone to F-35s and aircraft carriers and Virginia-class submarines, and frankly, we still need some of them, especially Virginia-class submarines, that's a tough transition to make. Even in a world where everybody rationally knows that we need more of these low-cost precise mass systems to increase the probability that the United States can prevail on a modern battlefield and to stay on the edge,

LINDSAY:
So Mike, how do you think about questions of force structure for this in terms of how the services spend their acquisition dollars? Let me take two examples. I've heard people argue that with advances in drone technologies having manned aircraft, think the F-35 are simply too expensive and are not a good investment and we should move away from that. I've talked to a number of Air Force pilots who tell me that's the craziest idea. Likewise, we talk about the United States Navy. You mentioned aircraft carriers. I've been in meetings with people who've referred to aircraft carriers as targets. How do you think through this, or how should we think through these questions, because the consequences, the significance of the results are massive?

HOROWITZ:
Those are really good examples. I think that where we need to go is what I would call a high-low mix for the future of the U.S. military. Essentially, small numbers of exquisite, more survivable, more stealthy capabilities. Think bombers, submarines, etc, etc, combined with really large... That's the high, really large numbers at the low of these precise mass systems, attributable, replaceable kinds of systems. And obviously then from an acquisition perspective, you would take different levels of risk for those different kinds of capabilities. For the capabilities at the high end, you need to make sure that they work because you are going to, as we have for the last several decades, plan to keep them for multiple generations. Whereas things at the low end that have lower cost points, you're going to plan to replace them pretty quickly anyway, which means you can, frankly, take more risk in developing and fielding them because if one thing doesn't work out, you will have another thing.
And on airplanes in particular, I think it is simultaneously true that much larger chunks of the Air Force could be more autonomous or more remotely piloted than it is today, and that the F-35 is not going anywhere in the short-term. I think the Air Force's roadmap for what they're calling collaborative combat aircraft, which are, think a next-generation UAV that unlike, say, a Predator and a Reaper used in the war on terrorism, is a little stealthier, can carry more weapons, would be more survivable in a conflict between the United States and China. The idea is you have several of them working with an F-35 or another American aircraft, to be able to then range forward and help project force or help project sensors on the battlefield. It might be that that ends up being a transition state and that as autonomy technology improves and as AI improves, you can move the human further and further back. But right now, whether we like it or not, the F-35 is not going anywhere.

LINDSAY:
This takes us back to your point, Mike, that success is going to be determined by which great power is able to determine how best to use new technology to marry it to existing platforms and having the right balance. But you're only going to really know the result of that if you actually get to a major power war, which none of us want.

HOROWITZ:
Hopefully we don't.

LINDSAY:
The point of all this is to create forces that deter the outbreak of war. So where does that leave us off?

HOROWITZ:
It's a really different way of thinking about it, in some ways. And I think that this is actually maybe one of the larger mindset shifts that I think a military like the United States is going to have to undergo. And by that, I mean the symbol of American naval power, say, since World War II has been the aircraft carrier in these larger ships, A. And as the cost of them has gone up and the time to construct them has gone up, the Navy has continued to get smaller. And this Trump administration, like we did in the Biden administration, like the previous Trump administration, is throwing billions and billions of dollars into different approaches to modernize the submarine industrial base and the ship industrial base to try to improve both our ability to produce those ships and our ability to maintain those ships.
But that's very different than a world where ... I mentioned collaborative combat aircraft for the Air Force before. A thing the U.S. Navy should definitively do is build, essentially a class of mini destroyers that, just like collaborative combat aircraft yoked together by an F-35, would be additional arsenals, essentially, of missiles and of sensors connected by something like an Aegis cruiser. That will then raise new questions about, "Would anybody be deterred by a mini destroyer going around?" Well, maybe not if they compare it to an aircraft carrier, but once they see what these things can do, I think that the environment will change. And that in some ways, the degree of confidence that we display will help influence how both adversaries and allies think about these platforms. But the one thing we decidedly don't want is a world where we emphasize deploying systems that we think would be good for deterrence purposes, but which we have no real confidence in if war comes. And so we need to figure out a way to better message mass in a new era of deterrence, rather than just relying on this small number of exquisite systems.

LINDSAY:
On that note, I'll close up this episode of The President's Inbox. My guest has been Michael Horowitz, the Richard Perry Professor of Political Science at the University of Pennsylvania and a Senior Fellow here at the Council. Mike, as always, illuminating to chat with you.

HOROWITZ:
Thanks for having me. Really appreciate it.

LINDSAY:
Today's episode was produced by Justin Schuster, Molly McAnany, Markus Zakaria, and director of video, Jeremy Sherlick. Production assistance was provided by Oscar Berry, Jorge Flores, and Kaleah Haddock.