Transcript

Remarks by Secretary Esper at National Security Commission on Artificial Intelligence Public Conference

Nov. 5, 2019
Secretary of Defense Mark T. Esper and Katharina McFarland, Moderator

(APPLAUSE)

SECRETARY OF DEFENSE MARK T. ESPER: Well, everybody settled down. We've still got conversations going on, huh. Must be some good issues out there.

Good afternoon, and thank you, Katharina, for that introduction.

You know, the work this commission is doing in bringing together academia, defense and business is critically important.

I'm going to scooch around here because I can't see over here.

So thank you for inviting me to speak today. It's really great to be here. The world around us is changing at a pace faster than ever before. New technologies are emerging that are fundamentally altering how we think about, plan and prepare for war.

Twenty-eight years ago I saw firsthand the transformative power of technology during Operation Desert Storm. As some of you know, I was a young infantry officer with the 101st Airborne Division. I took part in what became the deepest air assault into enemy territory at that point in history. And in only 96 hours, the 101st moved three brigades over 350 miles, cutting off the Republican Guard.

The Gulf War was the proving ground for a new generation of military weapons and equipment, from laser-guided smart bombs to stealth aircraft, to the first widespread use of GPS.

By liberating Kuwait and defeating the Iraqi military in a matter of days, American forces demonstrated our mastery of the digital revolution, and rendered what was then cutting-edge Soviet technology obsolete.

Our adversaries took note. And since then they've been trying to catch up.

Five years ago they surprised the world with how far they'd come. On July 11, 2014, Ukrainian forces assembled about five miles from the Russian border in southeastern Ukraine. Coming off recent successes against Russian-backed forces, the Ukrainian battalions were eagerly preparing a final push to the border. Suddenly they noticed the hum of Russian UAVs [unmanned aerial vehicles] overhead, followed by cyberattacks against their command and control and communication systems immediately, after a flurry of Russian artillery rained down on then. The whole episode lasted just a few minutes, but it inflicted tremendous damage. Dozens of soldiers were killed, hundreds more were wounded; most of their armored vehicles were destroyed. The Ukrainian offensive came to a devastating halt, all in a matter of minutes.

The world was quickly awakened to a new era of warfare advanced by the Russians. It's clear the threats of tomorrow are no longer the ones we have faced and defeated in the past. That is why our National Defense Strategy hinges on the ability of our forces to adapt to a security environment characterized by new threats from our strategic adversaries.

We are committed to making the investments necessary to accelerate our innovation in technologies that will help us stay ahead of the curve, especially artificial intelligence. Advances in AI have the potential to change the character of warfare for generations to come. Whichever nation harnesses AI first will have a decisive advantage on the battlefield for many, many years. We have to get there first.

Future wars will be fought not just on the land and in the sea, as they have for thousands of years, or in the air, as they have for the past century, but also in outer space and cyberspace, in unprecedented ways. AI has the potential to transform warfare in all of these domains.

The NDS remains the department's guidepost as we adapt the force to this new environment. The NDS prioritizes China first and Russia second as we transition into this era of great power competition. Beijing has made it abundantly clear that it intends to be the world leader in AI by 2030.

President Xi has said that China must, quote, "ensure that our country marches in the front ranks when it comes to theoretical research and this important area of AI and occupies the high ground in critical and core AI technologies."

For instance, improvements in AI enable more capable and cost-effective autonomous vehicles. The Chinese People's Liberation Army is moving aggressively to deploy them across many warfighting domains. While the U.S. faces a mighty task in transitioning the world's most advanced military to new AI-enabled systems, China believes it can leapfrog our current technology and go straight to the next generation.

In addition to developing conventional systems, for example, Beijing is investing in low cost, long-range, autonomous and unmanned submarines, which it believes can be a cost-effective counter to American naval power.

As we speak, the Chinese government is already exporting some of the most advanced military aerial drones to the Middle East, as it prepares to export its next generation stealth UAVs when those come online. 

In addition, Chinese weapons manufacturers are selling drones advertised as capable of full autonomy, including the ability to conduct lethal targeted strikes.

There's also ample evidence that China's developing and deploying AI to strengthen its authoritarian grip over its people. All signs point to the construction of a 21st century surveillance state designed to censor speech and deny basic human rights on an unprecedented scale.

Look no further than its use of surveillance to systematically repress more than a million Muslim Uyghurs. Beijing has all the power and tools it needs to coerce Chinese industry and academia into supporting its government led efforts.

Equally troubling are the outside firms, or multinational corporations, that are inadvertently or tacitly providing the technology or research behind China's unethical use of AI. 

Cooperation with Beijing has consequences, not just for democracy and human rights but also for the strength of our partnerships abroad.

If our allies and partners turn to Chinese 5G platforms, for example, it will inject serious risk into our communication and intelligence sharing capabilities. Our collective security must not be diminished by a short and narrow sighted focus on economic opportunity.

Russia has made its intentions equally clear, calling AI the future of humanitari – of humanity and describing the technology as the key to supremacy on the world stage. Moscow has already demonstrated its eagerness to use the latest technologies against democratic nations and the ideals of free and open societies. We shouldn't doubt their abilities on the battlefield, either.

I mentioned the Ukraine example earlier and we expect Russia to continue to deploy increasingly high tech AI capabilities in current and future combat zones. The United States, on the other hand, will offer a vision of AI that upholds American values and protects our fundamental belief in liberty and human rights.

We will harness the potential of AI to create a force fit for our time. We believe there's tremendous opportunity to enhance a wide range of the department's capabilities, from the back office to the front line, and we will do this while being recognized as the world leader in military ethics by developing principles for using AI in a lawful and ethical manner.

In line with the NDS, we stood up the Joint Artificial Intelligence Center, the JAIC. Its role is to integrate the power of AI across the many levels of the Department of Defense. Not only are we doing this in areas such as predictive maintenance and cyber defense, but also with more complex applications, like joint warfighting.

We don't approach AI, or any technology for that matter, as a panacea. We also see it as a tool to free up valuable resources and manpower so our war fighters and our operators can focus on higher priority tasks in a more efficient and more effective manner.

Our ultimate goal is to get the war fighter into the cloud. We must be able to pool our vast streams of data and deliver AI capability out to the tactical edge. This will require the wholesale commitment to modernizing our warfighting systems, cultivating a premiere workforce, and strengthening our partnership across the entire sector. We recognize these challenges, and we are committed to addressing them.

Our success is also contingent upon predictable, adequate and timely funding from Congress. The ongoing continuing resolution harms military readiness and impacts our ability to accelerate AI development at the speed and scale necessary to stay ahead. Our adversaries are not slowing down, and the United States cannot afford to, either. Congress must understand that short-term budget uncertainty has long-term strategic implications for our nation's security.

Now while technology is constantly changing, our commitment to the law, to ethics and to duty does not. The department's history clearly demonstrates our ability to invest in, develop and deploy systems that reduce risk to our war fighters while increasing our combat effectiveness for the ultimate purpose of protecting the security of the American people. We will ensure that we develop this technology in ways that uphold our values and advance security, peace and stability at the same time.

Some in the private sector have raised concerns about working on AI with the United States military. Unlike some parts of the world, American corporations have a choice in who they work with. That is the virtue of our free enterprise system.

But let me be – be clear, the question is not where I – AI will be used by militaries around the world; it will be. The real question is whether we let authoritarian governments dominate AI, and by extension the battlefield, or whether industry, the United States military and our partners can work together to lead the world in responsible AI research and application.

When America unleashes its collective genius of industry and government and academia, there is no one that can compete with us. I saw this firsthand in the Gulf War and our history is rife with other examples. During World War II, the titans of industry and hard-working patriots answered the call and transformed Detroit into the arsenal of democracy. After the Sputnik launch, we rallied our best and brightest, we created DARPA [Defense Advanced Research Projects Agency] and NASA and we took control of the space race. Mastering AI will require similar vision, ambition and commitment.

You and I are no stranger to these sorts of challenges. America has risen to the task before and we must do so again, but we need your help. We need the full force of American intellect and ingenuity, working in harmony across the public and private sectors. We need your leadership and your vision to ensure we maintain a strategic edge, and we need forums and commissions such as these to pioneer solutions that will deter aggression and provide for our collective security.

Thank you for your time and I look forward to our discussion. Thank you.

(APPLAUSE)

KATHARINA MCFARLAND: Thank you, Secretary, for the thoughtful remarks. It's clear that you and the department have been thinking about AI and what it offers to our military. Also, how it could be enabling our adversaries who may not share our values for our new and different (inaudible).

So I'd like to pull upon some of your comments, and thank you so much for sharing with us. I'd like to understand how the DOD might be communicating with industry and challenging them to solve our most pressing national security issues.

SEC. ESPER: Well we – we're reaching out in a number of different ways, everything from the traditional way of posting notices and RFPs [request for proposals] and things like that, to forums, to for a, to think tank sessions, to reaching out to academics directly, if you will.

You and I were talking beforehand. When I was secretary of the Army, we stood up the AI Task Force at Carnegie Mellon University in Pittsburgh and I – I spent a whole day out there working with some of the researchers and many of our industry partners that showed up for the – for the groundbreaking.

So we're trying to reach out at a number of different levels. We need to reach out with regard to industry and not – not just to the – the big players on the block, but all the way down to the small innovators. That's where you tend to find your greatest innovation, your ingenuity, and we need to make sure we do it comprehensively.

We – we – we've got to tap the best and brightest from across the country, again from all of those different sectors, and make sure we can – we can get to the end state quicker than the Chinese and the Russians can.

MS. MCFARLAND: God bless you. I think those things are – (Laughter.) – we have had a lot of conversation about speed; and I think this is an important conversation to expand on; and I understand you just mentioned the think tanks, that you've had the Defense Innovative (Innovation) Board just come in and provide some recommended principles for the ethics, because it's not just speed, it's conforming to our values as a society.

What are your thoughts about the DIB report, and have you had a chance to think about where these principles might be implemented across the department?

SEC. ESPER: Sure; well, as I said in my remarks, we have to, as we always have, conduct ourselves ethically and legally and morally. And I was very pleased by what I saw in the DIB report; it just came out last week and it listed a set of principles.

But in terms of going about and applying them to AI, it went back and reaffirmed the – the same principles we applied to other systems that we've – we've been using for many, many years. So I'm very pleased with the outcome of the report.

For those of you who had a chance to look at it, I think it – it – it's very comprehensive, it balances out a number of different things, it talks about the – the need for continued exploration of these topics to make sure we get it right. This is one we cannot afford to get wrong.

MS. MCFARLAND: Yes and – and one of the things that I'm very concerned about, and the commission has had a lot of dialogue about, which is our human resources, and how are we going to attract that talent and institutionalize it into the department.

Have you had a chance to think about those concerns, and what you might be able to do to attract the right type of talent to be able to do this business in the future?

SEC. ESPER: Yeah, I would say that, you know, with most things, talent is the key, is making sure you're – you're able to access the best and the brightest. You've got to be able to recruit them and retain them and keep them happy and busy and, you know, we've faced the same challenge over the past many years with cyber.

And I – I saw this not just from the government perspective, where the Army built the, you know, Army Cyber and – and started recruiting and retaining there, but also from the private sector, where we seem to all be – we all being industry and government and the academia – all competing for the same handful of people. And – because the – these are very talented, exceptional folks, and they – they – they have great opportunities to work in the private sector at – for large sums of money and do those types of things.

So what we have to do is make sure that we find different ways to attract them because we cannot compete with the private sector when it comes to compensation, but we can offer you the chance to serve your country, to do things that are very, very interesting, maybe do things that aren't legal in the private sector – (laughter) – and – but exciting nonetheless. But it's – it is a tremendous part. We – we tend to bring together a – a great deal of folks. And what I'll – I've always enjoyed with my time in service, whether it's in the military or in DOD, you – you've – you work around a great group of people who are focused on something bigger than themselves, bigger than the bottom line, and you get committed to that.

And again, I think Рit Рit Рit seems clich̩ all the time, but this is the space race. I mean, whoever gets there first is going to Рis going to dominate and we Рwe got to Рset aside Sputnik; we largely got to space first with Рwith what you needed and we dominated the heavens for Рfor Рand still do.

And we need to get there first on AI, and then maintain that lead, so it's going to be continuing investment. And so what we're trying to do now is make sure we leverage authorities that were given to us by Congress to – to make sure we can bring in people, we can recruit them, we can use different techniques to bring them in mid-career, we can bring them in with a different compensation packages and whatnot.

So we're – we're looking for ways to get outside of our own bureaucratic methods to make sure we can balance all of these things out.

MS. MCFARLAND: Super. It's so good that you come from a service background and roll up into this position. I think one of the things that we are seeing as a commission is the change of warfighting. And what do you see, now that you're in this position – coming from the Army with all of its challenges in the AI battle – now that you're in the secretary's position, what do you see that DOD is going to face in the future that's different with the way AI will implement?

SEC. ESPER: Well you – I – I think I mentioned this from my remarks. AI won't change the nature of war, but it'll change the character of war, which – which is – which is a – a major leap forward, if it can happen, and AI – AI will transcend everything we do. So it's not just warfighting, but it's going to be predictive maintenance, which is one of the areas in the Army, at least, that we're trying to get AI involved immediately.

You can think about what it does if you can really use AI to optimize your maintenance, you – you get higher reliability rates, you get fewer breakdowns, you get better efficiency on the system, et cetera. But then there's the other end, the warfighting end, and we talk about speed and decision making.

You know, these days, if you're – I – I wasn't in the armor; I was in the infantry, but I know a little bit about it, but you know, if you're a tank platoon leader or a tank commander, if you will, you're – you've got people actually looking out on the battlefield and calling out the enemy – enemy targets, if you will, enemy tanks, and you have sensors, too.

But imagine a world, though, where you have AI integrating into all of your sensors and everything, where the AI is constantly scanning a horizon and it's immediately, within milliseconds, it's – it's sorting out what is – what's a – what's a civilian truck and what's an enemy combatant vehicle, what's a tank and what's a fighting vehicle, which one has its turret pointed to you and which one doesn't, which one is your immediate threat?

And it can do that so much quicker, it could slew your gun, and it just allows you so much greater reaction to the enemy; and you can – and – and then that's where the man in the loop comes to decide whether you pull the trigger or not.

And that's how we've got to think about AI; enabling quicker, faster decisions that allow us to be successful on the battlefield and – and – and bring our folks home, too. And that's just a warfighting application. We can go toward anything. I mean, whether it's – heck, if it's even doing audits of the DOD.

MS. MCFARLAND: Yeah.

SEC. ESPER: Which has never been done before.

(Laughter.)

We're getting there. But – but everything. I mean, AI runs through everything we do and we've got to be able to make sure we get it – get it. That's why we're trying to move as quickly as we can to the cloud.

MS. MCFARLAND: Excellent. You mentioned during your preliminary notes here about your interest and your energy behind AI. And when you actually went through your confirmation hearing, you mentioned it as one of your highest priorities.

SEC. ESPER: Yeah.

MS. MCFARLAND: And you referred to the Joint Artificial Intelligence Commission – Committee. What are your thoughts for the future, and how do you think your leadership will be able to top down, as well as bottom up draw our department past that bureaucracy that you're discussing into that future?

SEC. ESPER: Yeah, well the – you know, the acquisition system is not – is not as efficient as it should be. We're ...

MS. MCFARLAND: Yes, sir.

SEC. ESPER: We're – we're trying to take advantage of laws given to us, and authorities by Congress, but it's slow. And the biggest problem with DOD in terms of acquisition is – is the culture, right? It's very risk adverse.

So we've got to change the culture. That'll take time. You know, you change the laws first, then you change the regulations and the practices. And I think the services are moving forward on that at different rates. But we've got to empower the JAIC to – to be able to cut through these things to make sure we get there quicker.

Again, we're – we're in a race, we have to get to the end state quicker than the Chinese can, quicker than the Russians can. And there are a few key technologies out there. I put AI at number one. You know, two, three and four look like directed energy, and hypersonics, and a few other things like that. But at – even with those systems, whether it's hypersonics, directed energy, AI is still going to enable them in – in terms of how you employ them, how you maintain them, all that. So that's why AI to me pops up as number one.

MS. MCFARLAND: You mentioned in your conversations here the issue of bringing the best and brightest. Have you seen some good collaboration between the government, academia and industry? I recall – or at least I seem to recall when I was working with you in the Army, there was some activities.

Have you got some examples that we can bring to the forefront to get people past that risk equation?

SEC. ESPER: Yeah, it's a good question. I – I – you know, I wish I had some at hand. I – I mentioned before the – we were – we kicked off a AI Task Force at Carnegie Mellon University, and we're doing some good initial work there and had a – a number of different players involved.

Obviously, the Army is deeply integrated in Austin, Texas, with the Army Futures Command. They're doing a lot of work trying to cut through bureaucracy, get straight using the cross-functional teams to do that. But again, I think all of the services are looking for different ways where we can – we can really accelerate the progress, because we need to get there, we need to make sufficient investments; and right now DOD is in the final weeks, if you will, as – as you recall, of building its budget for the next year.

And – and again, AI is one of those core, critical technologies we need to get to; but – but it's not just – it's not just the money, it's the people. I probably put people number one, and then all of the systems that you need to enable them to do their jobs and do them well.

MS. MCFARLAND: Yeah. You mentioned the talent and trying to recruit it. There's a pipeline, as you know, that you went and experienced, not just in the Army, but now in the broader context of DOD, where science and technology, early research and that investment is so important.

What are your thoughts about the future in that investment and where do you think it might be needed in the department?

SEC. ESPER: With regard to the – on the personnel side?

MS. MCFARLAND: Yes.

SEC. ESPER: Well you know, the – the biggest thing – one of the biggest things we were pushing in the Army and – and I'm confident that Secretary McCarthy and – and – and Chief of Staff McConville still are, is talent management.

It's – you know, we had to overhaul the – the personnel system ...

MS. MCFARLAND: Yes.

SEC. ESPER: ... because it was just holding us back in too many ways. And it's – and that's just for the military side. You know, there's the – the – the civilian side, as well, that needs looked at, too. So, DOD does great work. We – you know, we do a number of things really well, but when it comes to talent, we're still working in a industrial age system.

MS. MCFARLAND: Yeah.

SEC. ESPER: And it's – you know, it's regulated by – by – by the Executive Branch; it's regulated by Congress; it has a number of constraints put on it. But we've got to get beyond that, and just be able to think outside the box, because again, at the end of the day, it's the talent.

And it's not like we're – we're in a non-competitive environment. We're competing against industry, we're competing against think tanks. All – you know, these – these folks are in high demand, and they're a – a low density, at least, pool of talent at this point in time.

MS. MCFARLAND: Yeah. And you have an opportunity here with this community, it's a mix of academic, it's a mix of civilian, et cetera. Is there anything that you would call for them to think about that would help you in your problem solving?

SEC. ESPER: You know, I – I – I think the more that you can help us point out what are – what are the obstacles we are putting in our own way, right? We have enough challenges out there external to us, exogenous. So how do we – what do we need to do better, what are you seeing that I don't see?

I – I try every month to meet with – with groups of CEOs or heads of associations and talk about what can we do better, how can we see ourselves better? Because typically, what I hear is everything is okay, everything is great, everything's green, no problem. But it – when you reach out there, you go walk around and visit folks and talk to companies and talk to entrepreneurs, you get a different story.

And we're trying to just beat those down one at a time as we – as we realize, oh, we're not doing this well or we could – if we make some adjustments to the system here, we could really up our talent. So we're looking at – I'll – I'll – you know, we'll – I'll take any problem and it allows us – if it can allow us to do better or – or, you know, ideas you have for us to perform, all those things help because this is too important ...

MS. MCFARLAND: Thank you.

SEC. ESPER: ... just to – to treat as anything else.

MS. MCFARLAND: Yeah. Sir, your remarks and your inputs have been excellent. Is there anything that you would like to share beyond what I've asked you? Is there any thoughts that you would like to have for our folks here?

SEC. ESPER: No, I – well, yes, I guess, I – I – I'll just keep foot stomping it, is DOD doesn't – doesn't have the monopoly on great ideas here or – or certainly all of the talent. So much of it is coming from the private sector.

So we really need your help and – and be cautious of – of what's happening out there in the world. As – as I like to tell our NATO allies and European friends – and I was just in Brussels two weeks ago talking to the German Marshall Fund, on this topic, by the way. I tell them, 'don't – don't write off what we're saying as United States scaremongering, or in this case DOD scaremongering about China. Don't think we're over – overstating the problem.'

There are serious issues out there, and we've been asleep at the switch now for quite some time/ And we're finally waking up here in the past couple of years. And the National Defense Strategy is what – what has pointed DOD in the right direction, saying we're now – now in an era of great power competition; China is our greatest strategic competitor number one; Russia number two. And we need to be prepared for high intensity conflict across five domains of warfare – five domains, no longer three, and that's where we're headed, and – and I – we need to get your help to get there, so ...

(CROSSTALK)

MS. MCFARLAND: Thank you sir, so much, very much. 

SEC. ESPER: Thank you. Thank you.

MS. MCFARLAND: Please, thank the secretary.

(APPLAUSE)