The vital race for cutting-edge technology

In the following edited transcript from proceedings at ASPI’s Sydney Dialogue, Michael Pezzullo, secretary of the Department of Home Affairs, interviews Eric Schmidt, chair of the Special Competitive Studies Project, co-founder of Schmidt Futures and former CEO of Google, on geostrategic technology competition and why it matters.

Discussing the crossover between technology and security issues, they dive deeply into this competition, including whether the US$18 billion to build an aircraft carrier might be better spent on a mass of much cheaper drones and whether air forces need crewed or uncrewed combat aircraft.

They examine China as an autocratic competitor to the West that is run by technocrats and is very capable of inventing a new future for the whole planet. They note that China is leading already in vital areas including communications, surveillance and financial services and that it’s on its way to dominating new energy as in the infrastructure of electronic vehicles.

Michael Pezzullo: There could be no more important topic in the strategic area of thinking and policy at the moment than the crossovers between technology and security, and what a splendid program ASPI has put together. Eric, you’ve got a very important Foreign Affairs article in the current edition, on innovation power. And late last year you did some work with former US deputy secretary of defense Bob Work on a related issue of the Offset-X strategy.

You spent a lot of time in technology, you’re a tech entrepreneur, you’re globally famous. At what point in that intellectual and personal journey did you start to think about these security issues, which you have so been focused on in most recent years?

Eric Schmidt: Like everybody else in tech, I was obsessed with myself and what tech could do, and I thought we would take over the world. And I think my first comeuppance was when we bought YouTube, and understanding that the world did not work quite the way we thought it was going to work. A few years later, Ash Carter, who was the secretary of defense, and unfortunately just passed away of a heart attack, asked me to serve as a consultant to the US military, which I thought was curious because I’ve never been a military person. But I became very interested in national security and what technology was going to do.

That led me to work on both the AI aspects of national security, but also what, in my view, is wrong with innovation, which ultimately led me to the view that there’s a difference between hard power and soft power and innovation power. I fundamentally believe that the future will be driven by innovation and not by hard power and soft power. So, I guess I’ve ended up at a uniquely different point of view than I thought that I would end up with.

Pezzullo: Well, the blending of perspectives, I think itself creates new possibilities and new ways of seeing things. Your very deep expertise in tech, as I think you just said, that you never came up as a security guy or a military guy, I think was your reference. But the fact that the late Secretary Carter tapped you on the shoulder to come onto that process, I think speaks volume to his own vision and innovation.

When I came into government service 36 years ago—and I know that, if you can see me, you’ll say, ‘That can’t possibly be right because you couldn’t possibly be that old’—the most I could tell my mother about the work that I was doing in the defence intelligence community was to refer her to The Hunt for Red October and say that this technology that the Soviet Navy was demonstrating, as Sean Connery famously said, ‘Tonight we sail into history.’

What strikes me about that period as a young intelligence analyst was that technology was all within government. It was very government-centric. The thought that there were technology companies who could give us capabilities that were not available otherwise to military intelligence and security services was baffling to us. And indeed, in that late competition in the Cold War between the US and the Soviet Union, one of the key factors that led to the breaking of the Soviet Union was that they couldn’t keep up with the dynamism and the innovation of the US economy. Which was sitting off to one side producing great consumer outcomes and great benefits for its society.

And the military was able to actually develop technology in its own ecosystem, for instance to hunt submarines. That paradigm over the 36 years that I’ve been in this business has completely flipped. And indeed, most of the technologies that you speak about in some of your recent writings, such as swarm drones or the ability to upload sensitive government data into secure clouds, so that if your physical ministries are attacked, for instance, you’ve got a resilient backup of your data. All of that’s been generated out of the private sector.

For the benefit of those very few in the audience who would not have read your Foreign Affairs article, one of the most interesting insights that came out at me was your reference to, as human innovation traditionally has occurred, it hasn’t, in a sense, catalysed further information in and of itself. You’ve got a memorable line in the article about faster planes never begat even faster planes. There was a whole new step-change in plane technology. Where things like artificial intelligence, particularly generative AI, and one day perhaps AGI, will itself generate not only its own advances but will establish advances or power advances in other like fields. Can you draw out some of the further thoughts that you’ve put into that article on innovation power?

Schmidt: Well, a simple model for everybody is hard power, we understand that, that’s military strength, we’re in charge, you’re not, kind of thing. Soft power, of course is economic power, cultural power, Hollywood, values, Western values, all that kind of stuff. We all understand this. The difference now is that we’re locked in a battle with China that will define the future for the rest of our lives. And China is not like the Soviet Union. At its height, the Soviet Union was only a third of the scale of the United States, and although they were certainly dangerous and they’re dangerous now, they were not fundamentally a global platform competitor that could affect the West. But instead, now we have an autarkic, if you will, autocratic competitor that is run by technocrats, that is very capable of inventing a new future.

And they can invent a new communications feature, that’s called Huawei and 5G, which is also known as a signals intelligence nightmare. They can invent a new application known as TikTok, which has taken over all the teenagers in the world. Now, I am perfectly happy that somebody knows where our teenagers are because the average family doesn’t and the Chinese clearly do. But the principle of China being able to do first, communications, and second, apps, is something that, if you’d asked me five years ago, I would’ve said not. They’re on their way to dominating new energy that is the infrastructure of electronic vehicles. They’re already ahead in things like surveillance and in financial services. What’s next?

We need to get our act together. Here we have these great people in Australia, great people in the UK, great people in the US, doing AUKUS. What are you doing to make sure we stay ahead of what the Chinese are doing? That’s true in AI, it’s true in quantum, it’s true in material science, it’s true in synthetic biology, it’s true in various forms of transportation and so forth. Now, I’m going to argue that the way the world works is there are these platforms and the platforms become global platforms, and I want the platforms to be built in the West, right? I want them to reflect Western values. Can you imagine if the internet had been built using Chinese principles? Everything would be surveillance all day; that’s clearly not good.

One final comment is that I went to Ukraine because I was interested in what happens when the tech people actually work with the military. And to see the way in which tech people would solve the problems of communications and targeting and surveillance and so forth, using drones and many other navigational things, is remarkable. And they do it at one-fiftieth, one-fiftieth of the cost of the military systems in the West. I’m so glad we’re at peace. We can waste all this money on really big things instead of building a large number of little things that work together. That transition is the transition that all of us, Australia included, have got to get through, and you’ve got to get through it quicker than we are, in my view.

Pezzullo: I want to come back to some thoughts you’ve developed around prescriptions for government, including through the work you did on the National Security Commission on Artificial Intelligence and the work that you’re sponsoring and chairing on special competitive strategies, of course, which really speak to the issues that you’ve been focused on in recent years.

You touched on at the very end of your remarks about Ukraine that our speed to decision, our speed to innovate and now speed to adapt needs to be completely deconstructed.

You came out of Silicon Valley, that was built on disruption. It was built on disrupting your competitor, coming up with something that had an edge, coming up with something that no one had thought of before. Bureaucracies, and I touched on this in terms of my own time as a government official for 36 years, very much have got the opposite DNA. Set procedure, predictability, the management of risks so as to account for the expenditure of public dollars.

How do we bring these two different perspectives together? Because on the one hand, obviously our democracies are reliant on the rule of law and the separation of powers and the oversight of officials such as me, in terms of expending hard taxpayer-won dollars. You did some work on the Defense Innovation Board for instance. How do you start to break down that mentality, keeping our fundamental democratic oversight principles and our democratic accountability principles in place? How do you start to change the thinking in public bureaucracy such as the one that I’ve served in for 36 years?

Schmidt: In the first place, the people in all the positions are trying to do their best. The problem is that they have the wrong formulation of how to win. I’ll give you an example. From a technological perspective, in terms of military achievements, there’s no place to hide anymore. Everybody has various forms of commercial and military satellites. You can’t hide your aircraft carrier anymore. You cannot hire your big launcher thing. If you look at the HIMARS, for example, in Ukraine, they’re very interesting because there’s a tree, you park it under the tree, nobody can see it. Then you move it, you shoot it and then you move it somewhere else and you park it under a tree again. So, we have a wrong conception of military power in that things will be quick, decentralised, very fast moving, hide, in and out. That breaks almost every military doctrine, starting with aircraft carriers and you can go on from there. And yet we still build them.

I happen to love aircraft carriers because I like big ships and they’re great fun to visit, and I’ve enjoyed them, but the US$18 billion that the latest carrier cost could be spent, think of the number of drones that you could build. So, what happens is that the bureaucracy, because it gets captured by the contractors and by the dominant thinking of the time, does not actually ask the question, ‘Will it work?’ There are many, many examples, on the US side, of total failure. A good example is the LCS, which was the littoral combat ship which was designed to fight battles near the beach. And the conception did not take for granted that there would be missiles, they could knock it out close to the beach. It’s just an error.

So, the way you solve these problems is you have a rule that you have a choice A and a choice B, and they’re different, and you compete them against each other. A simple proposal for Australia is take some of the ideas that I’ve been working on in Offset-X and try them in little groups, and just say, ‘Do they work?’ For example, Australia’s surrounded by water as we all know, and your battleground, if you will, is the Pacific. Where are your underwater drones—the militarised underwater drones? How good are they? How far are they? Let’s run that and compete that against the next big ship you’re doing, right? Have an honest competition because I can assure you that that drone is a lot cheaper.

In America there’s a discussion about the next-generation fighter. Now, it’s obvious to people like me, and I’m a private jet pilot as well, that the next version of these fighters should be automated. That is, it should be a drone, right? Because we can do that better and there’s all sorts of reasons. And yet the military says, ‘We have to have a pilot in the air as well as some drones.’ Okay, well let’s compete that. Let’s have the pure drone one against the military one. And by the way, the one involving humans is going to cost 10 times more because humans are important and we don’t want to lose them, and so forth and so on. So, the only way I think, just to be very blunt, that we can break this log jam is by hard competition of approaches.

Instead, what happens, in my experience with the US military, is that they set up really false tests. My best example is missile defence, which doesn’t work very well. What we do is once a year we arrange for a target to be shot up and another target very carefully to be shot so that they hit each other. Well, that’s a great way to prove that the thing can hit each other once a year if we know where it is, which is precisely not how missiles work. So, let’s have a real competition and look for some new choices. I think they’re there. And the key thing I will tell you is that there are plenty of citizens in the Western countries who want to work in companies to serve the nation as civilians, to make money, and to make our national security greater. There’s plenty of people who want to help.

Pezzullo: I was going to come back to talent, and you’ve given me a perfect segue at the very end of your remarks there. You used the phrase ‘Offset-X’, which is something that you have been working on, and you co-authored a piece with Bob Work just before Christmas in The Atlantic.

Of course, that builds on the previous offset strategies in the 1950s and ’60s where we talked about that highly militarised competition between the Soviet Union and the US. A refresh of the offset strategy in the 1970s, to thwart a massive Soviet ground attack into Western Europe that would otherwise threaten to go tactically nuclear very quickly. Bob Work, when he was at the Pentagon under Ash Carter of course introduced the third offset. Can you just spend a moment or two just unpacking your idea of Offset-X?

Schmidt: The current situation in military strategy is high precision, and that’s what the third offset really means. Precision is important because none of us want collateral damage; that’s obviously bad. We don’t want to harm anyone who’s not a combatant, and we want to deliver our weapon or our defensive system precisely. That’s what the third offset is. What we argue in Offset-X is that there’s another level of networked autonomous approaches where … And you want to think about swarms and you want to think about autonomy, you want to think about the use of AI, and you want to do it systematically. This concept, I’ll give you another example. I have no courage at all. I have great admiration for our military people because they actually have real courage. I have zero, but if I found myself in a tank or as a soldier, I want 50 drones ahead of me and 50 drones behind me. The 50 in front of me, ready to attack and watching out for me, and the ones in the rear making sure that I’m not approached by the rear.

Now today we have one drone for every 2,000 soldiers. I want 50, right? So it’s an inverted way of thinking about it, that we want a very small number of humans who are very heavily protected, who have extremely autonomous systems that are under their control in the battlefield. It’s almost impossible to win against an autonomous, decentralised, human-controlled network of the kind that I’m describing. You just can’t make enough progress. You can’t shoot enough satellites down because there’s more, you can’t shoot enough drones down because there’s more, and so forth.

If you look in Ukraine, right now they’re having sort of a World War I–type battle. If you look at the lines essentially below Bakhmut, near Zaporizhzhia and so forth, there to Crimea—and I went there to investigate, by the way, so I know a little bit about it, and these people are incredibly courageous—they’re locked in the World War I-style battle, tanks on one side, tanks on the other. The only solution for that is going to be drones that go up, over and around, and from the rear, which the Russians can’t defend against. Here we have a real life example of how you’re going to win, and I want Ukraine to win for all the obvious reasons.

I think what happens in the military is just—because we’re not, thank God, at real war in most of these countries—they’re not really doing authentic exercises. To me, you would say, ‘I have an asymmetric competitor coming up against Australia in the Pacific. That asymmetric competitor is using all these new tools. How am I going to defend myself?’

First, are my bases defended from drones? How do I hide my big assets from the satellites from my competitor or my enemy? These are questions that I would encourage you and the leadership of the country to have a hard reckoning of. What’s nice about Australia and the UK, by the way, is that because you’re—just being blunt—smaller, you also have better control over things. Whereas the US military can be understood as a large country of different fiefdoms which has many different bosses and many different incentive systems, including jobs programs and continuity programs, and state funding programs, and so forth, I hope. And I don’t think that Australia has that problem. You could make progress. In fact, you could show the US how to win.

Pezzullo: There’s something in the spirit of democracies that seems to me to be a natural advantage. If you could speak on that. Some of your writings speak to visa programs, what we’re doing in our universities. The entrepreneurship and the innovation that’s almost embedded in, if you like, capitalist, free-thinking societies. Then I’d like to ask you a provocation, as a devil’s advocate, about a contrary view that I would want to put, that goes to whether the world should in fact be on the path to decoupling.

Schmidt: Sure. I don’t understand where all this foreign opposition comes from within these countries. Don’t you want the top people in the world, in Australia and in the US, and in the UK, working for the national goals of those countries? Why would you not want the smartest Indians and Russians and Chinese out of those countries and in your country? It’s obvious if you think about it, right? Because these are uniquely talented people. And yet in America we have a situation where we take people, train them to the teeth in quantum and then we kick them out and they go to China and they create a quantum program, which is going to be used to decrypt and basically analyse all of our data when quantum finally works. It’s insane.

And if you talk to smart people around the world, they want to work with us. They want to work in your country or mine. Why? They’re nicer countries. They’re better run, they’re better places, they have a higher quality of living, their kids are happier. This is our advantage and yet we’re squandering it. It’s insane. Now the Chinese should be taken very seriously because their model, which is a different kind of model, is effective. So, what advantages do we have? We have the same weather, they have more cash. They have certainly as large a diaspora as we do in terms of English, but we have open and free immigration and people want to come to our countries and not to China.

Pezzullo: You shouldn’t assume that the premise of this question is one with which I agree, but what would you say to those who say, ‘Hang on, if you go down that path, the path that Eric Schmidt is promoting and representing in his writings and his thinking, you actually entrench a decoupling or a separation at a global level, at a strategic level, of effectively two systems, a Western US-led system and the Chinese system.’ And indeed, in one of your articles, you talk about a clash of systems, not just a clash of states. What do you say to those who would say that your prognosis, whether it’s right or not, the prescription of it actually entrenches separation and entrenches, if you like, decoupling to the detriment of global humanity?

Schmidt: I had hoped my argument was wrong. And when I started working on this, I actually hoped that you could get a group in China and a group in America, I’m just using it as a simple model, to agree on common platforms around important things. In each of the platforms that I studied, for one reason or another, they got split. They got split for tariff reasons, they got split for signals intelligence reasons, they got split for political reasons. There’s just no scenario, in my view, where the really sexy stuff, so the strategically powerful platforms, are going to be common. It’s not going to happen. There’s too much danger to each country. And by the way, the Chinese don’t want it any more than we want it. They don’t want our, what they view as democratic core values, in their country.

Neither side wants to make the marriage work, so I think a reasonable model is the following. I think we’re not decoupling in the sense that most people use it. We’re going to have a great deal of non-strategic partnering with China, and Australia will. There’ll be lots of trade in things that are relatively common—steel, oil, minerals. Australia’s rich with everything; there’s going to be lots of that. But there’s not going to be agreement on common standards about ships, common standards about operating systems, common standards about how the internet works. The systems are not … it’s just not going to happen.

And the result is, unfortunately we’re going to see roughly a China world, a Western world. And the key question is not will that happen, because I’m quite sure, but what happens for the countries that are trying to be both? If you think about it, to be obnoxious, Germany wanted the market of China and the security of the United States, and now they’ve lost the market of China because they wanted the energy of China, and they can’t get it anymore because of this horrific war.

It’s really tough for the countries that are not obviously in one alignment. I’m predicting that India will become more and more Western. It’s traditionally not aligned. I can’t tell about the Arab countries. I think they’re playing both sides pretty well right now. Russia and Belarus will clearly be rogue actors for the next 20 or 30 years because of their horrific war. We have opportunities in Africa and in Indonesia and so forth, which you all understand very well. But I think this is the truth right now, and we better get ourselves together so that we win in the competition around strategic platforms. I don’t want to be using Chinese operating systems to do my communications, I just don’t. I don’t trust them.

Pezzullo: We have two questioners.

Alan Duffy from Swinburne University: Dr Schmidt, you raised the opportunity that drones present where quantity really has a quality all of its own. Are we not risking the challenge by our adversaries owning the largest drone fleet, Chinese companies in particular DGI, where we are attacking them head-on exactly where they are strongest? Does that not run a risk that we are playing ourselves off against their strength?

Schmidt: Alan, that’s very well said. And the fact of the matter is that 90% of the commercially available consumer drones are built in China. That does not, however, mean that there are not opportunities to build very powerful national security drones. And I’m now aware of literally a hundred startups related to the Ukraine challenge, which are trying different approaches. As you know, in combat, the first thing you do is you block GPS. It’s the first thing that anybody does. I read with some interest, Qantas reported that one of their airplane’s GPS was jammed, and thank goodness there was no harm to that. So, there’s something going on with GPS.

The military drones are going to be quite different from the DGI drones, carry heavier payloads. What I believe will happen is eventually there’ll be a call for the equivalent of a giga-factory for drones. We’re just going to need a lot of them, built in the West with Western components.

Mike Woodrum: I’m Mike Woodrum from the University of Auckland and from Trans-Exxon, a brain-computer interface company. We are currently in a time when we’re clearly moving from natural and organisational intelligence as to the way we run the world, to a combination of natural, artificial and organisational intelligence. My question is, are we already past the Offset-X age, and how long is Offset-X thinking sustainable in a world where artificial intelligence is deeply integrated everywhere? And is the sort of competition you’re talking about, these bipolar or multipolar competitions among humans, even sustainable for even five years, let alone 30?

Schmidt: Well thank you, Mike. Offset-X is designed to be powered by AI. Essentially all the automation, all of the autonomy, all of the navigation is AI enabled. And one of the things that we say very clearly is that we want automatic weapons, but we want the humans to make the decision to deploy them. So, we’re very clear where the boundary is. I think that for at least 10, maybe 20 years, it’s going to be very important that any system that uses AI have a human not only in the loop but watching what it’s doing. Because as you know, the current systems, whether they’re diffusion models or transformer models, tend to make mistakes. They tend to hallucinate and so forth and so on.

I’ve been doing this in one form or another for 50 years. I’ve never seen the gains and the speed of what we’re seeing with the large language models now. My industry has gone insane. It also means that the rest of the world is going insane, too. So not only are we all going to invent this AI future, but so are the Chinese. We can’t prevent it. We can’t just hold it to the equivalent of Los Alamos and hold it to ourselves for a few years, for advantage. We’re in a race, and I’d say too, I think the most important thing to understand is that China is a new kind of competitor. They are a partner in the sense that we can rely on them for some things, but they’re a competitor in others. Notice I didn’t say ‘enemy’.

If you look at Russia, Russia is becoming legitimate as an enemy to the West with respect to their illegal activity in Ukraine, and so there’s a good test bed for us to prove that this new innovation model can actually help hold back the old model. I think that we need to seriously understand that here’s a real land war that’s going on in front of us. As Australians, how would your systems hold up under the kind of attack that the Russians have done? Let’s hope that you conclude well. Now let’s do the same analysis for China’s strategy, which is much more organised around hypersonics and drones and autonomy, and you might be not as happy with yourself. That should spur action. I’m very happy to help you guys and work on this. I care a lot about it, and I really, really am thankful that you invited me. And please have a good morning. It’s evening here in New York, but I look forward to being there soon.