How the Defense industry is playing catch-up in artificial intelligence

Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne

Most artificial intelligence innovation is taking place in the consumer sector. Perhaps not enough in the defense industry. But some new research shows how extensively the largest defense contractors are investing in AI startups to supplement their internal AI spending. To dig deeper, Federal Drive with Tom Temin talked with a researcher at Georgetown University’s Center for Security and Emerging Technology, Ngor Luong.

Interview transcript:

Tom Temin: Ms. Luong, good to have you in studio.

Ngor Luong: Thanks Tom. Thanks for having me.

Tom Temin: And what were you looking at here? The fact that there is just not enough investment in AI for national security needs or what’s going on here?

Ngor Luong: Yeah, so this report was looking into a couple of things. I think it’d be helpful to break this down into three parts. First, we study why artificial intelligence is important. The reason why we’re looking into artificial intelligence is because it is a disruptive emerging technology, and it also has dual use capabilities as well. And we also know that militaries around the world are viewing AI as critical to their country’s economic development, national security and societal welfare in general. And we also looked at the largest global defense companies. And the reason for this is because these are really significant suppliers of emerging technologies, such as AI to militaries around the world. And they are also important integrators of emerging technologies, cutting edge technologies, including AI, biotech, etc, into large defined systems and platforms. And so it’s really important for us to understand the reality of AI innovation within these major defense companies. So there are a couple ways for them to develop their AI capabilities. One approach that we can actually observe is their private equity investment activity in AI companies. And so in our report, not only that we track them, which is an acquisition activity, we also track their private equity, venture capital investment in AI companies. The reason why this is important is because they may be interested in investing in or acquiring AI companies to access external AI innovation, for a couple of reasons. One, they’re doing this because they want to remain competitive in the emerging tech market. As you know, tech giants like Microsoft, Google, etc, are leading in AI innovation. And so it’s very important for them to remain competitive, and to investment in M and A, actually, perhaps going to give them the advantage over competitors that only focus on AI in house R&D. And the last point is, when you look at the US Defense landscape, they potentially see themselves as a bridge between the DoD and the private sector. So all of these components together really help us understand the state of innovation within this major defense companies.

Tom Temin: And the companies you looked at are like Northrop Grumman and Lockheed, , those types of companies?

Ngor Luong: The largest, the 50 largest global defense companies.

Tom Temin: And what is the extent collectively of their investment in AI over the last, what couple of years?

Ngor Luong: The data tells us that few of the top global 50 defense companies are investing in acquiring AI companies, and specifically 11 of the 50 global defense companies are using this mechanism to acquire external AI innovation.

Tom Temin: And that struck you as low?

Ngor Luong: Exactly. 11 out of 50 companies. And in terms of M and A, there are only five defense companies that actually acquire six AI companies, we observe between 2013 and 2020.

Tom Temin: Yeah, so that’s a pretty low pace of acquisition and investment, then,

Ngor Luong: Absolutely.

Tom Temin: Do we know the dollar value added up?

Ngor Luong: Yeah. So they made a total of 52 investment transactions in AI companies over this period of time. And the total amount is around $1 billion in disclosed value. The reason why we said disclosed value is because some of the transactions, obviously they don’t disclose their transaction value. So we don’t observe those.

Tom Temin: Yeah, so in the grand scheme of things, it ain’t much really then is it?

Ngor Luong: Exactly.

Tom Temin: We’re speaking with Ngor Luong, she’s a research analyst at Georgetown University’s Center for Security and Emerging Technology. And what about the internal investments in growing natively within the companies of these 50 companies, their AI capabilities? Is it possible to see in to what they’re doing?

Ngor Luong: Yeah, so that is one mechanism that we can possibly look into, although we don’t have access to the internal data so it’s very difficult to see that that’s one gap in our research, we only look..

Tom Temin: You have to go to China to get that.

Ngor Luong: They’re actually tracking our internal R&D. So yeah, absolutely if we had the chance to do it. I think if these defense companies are actually thinking about looking for ways to improve their AI development, it could potentially track their own progress using their internal data.

Tom Temin: Right. So then is it the conclusion of your study that there is simply not enough investment going on in AI by the top 50 defense contractors to serve national security needs?

Ngor Luong: The conclusion is they are not relying much on this mechanism to access the external AI innovation that is being developed in the commercial sector. So there are different ways that could potentially do this. This is not to say that they’re not developing AI capabilities and capacity at all. This is to say that they are not relying much at a significant level on this mechanism.

Tom Temin: Because we know there are many, many startup in mid-sized companies specifically in that business. And you see Microsoft buy them from time to time, and maybe Google buys them or Alphabet buys them from time to time. So there’s competition for those assets, correct?

Ngor Luong: Yeah, absolutely. A lot of innovation is happening in the commercial sector, as you already list a couple of the tech giants that are doing a lot in terms of injecting capital flow into these companies and also investing in their own internal R&D. And yeah, there’s a lot of capital going around in the private sector. I think, if the defense companies are really serious about investing in AI and developing the AI capabilities, I think you should look in this direction.

Tom Temin: And looking at it from the point of view of the Defense Department, they do have mechanisms for getting at AI outside of those large companies, the Defense Innovation Unit, the joint Artificial Intelligence Center, or the JAIC. They are using, to a large extent other transaction authority to quickly buy technology from some of these startups in some of these innovative companies. So maybe the question is, the lack of investment by the large defense contractors may not net out that big of a problem for DoD, which has other ways to get the AI into the department.

Ngor Luong: We are interested in answering that question as well. We know that it’s a DoD, particularly the US Defense Innovation Unit, has been screening a lot of non traditional suppliers, such as AI startups, for their cutting edge technologies, right. But the issue is the defense contractors or the largest global defense companies are the ones that are doing integration and also supplying large equipments that are important to our existing large systems and defense.

Tom Temin: Right. So next generation tank is AI enabled, it’s probably best if it’s integrated by the builder of the tank, in other words.

Ngor Luong: Absolutely, absolutely.

Tom Temin: Because the AI requirements of military are a little bit more rigorous, say, than if you’re making five flavors of cookies, and you want to know what’s selling here and where and so I can direct my bakery operations to the current demand. That’s a very different AI need than a military application, which has to be auditable, has to be transparent, it has to be according to doctrine of the military, it can’t make its own decision, and so on. The requirements are much tougher, aren’t they?

Ngor Luong: Absolutely. And they also have the DoD in mind when they are developing these kind of technologies for our warfighting capabilities, and so on. They also know how to navigate our bureaucracy, understand the laws and regulations in order to actually successfully integrate those technologies into our existing systems.

Tom Temin: Alright, so what’s your bottom line advice?

Ngor Luong: So we’re not saying that these largest defense companies are not doing AI. We’re saying that one approach that we can actually look into shows that they are not using it as much, the mechanism is being underused. What they could potentially do in the future is to leverage the innovation that is being developed in the commercial sector. And another thing that I want to mention is one major difference that we noticed between the defense companies that actually invested in AI companies, and those that did not, is what are they actually have a venture capital subsidiaries, right. These subsidiares are basically those that can go out and scan any rapidly develop AI capabilities in the commercial sector. And these companies are huge. So having a division that can actually go and invest in AI companies can be really helpful. So in future if that is the direction that they can potentially explore, that would be great.

Tom Temin: And perhaps one reason of the sluggish pace of investment in these startups and innovative companies by the defense contractors is maybe they want to keep their sub contracting options open.

Ngor Luong: Yeah, we don’t know. There are a lot of ways that they couldn’t potentially be developing AI, right. They could probably invest in industrial research and development. They could be investing directly into universities or establishing subcontract relationships with other industry players for the militaries. So these are very important mechanisms that we should probably further explore, but that needs internal data, which the defense companies and DoD have access to, and we highly recommend them for they’re looking into this.

Tom Temin: Alright, well, you got your next research assignment them. Ngor Luong is a research analyst at Georgetown University’s Center for Security and Emerging Technology. Thanks so much for coming in.

Ngor Luong: Thanks for having me.

Related Stories