Greg Allen: Artificial Intelligence Offers Exciting Capabilities for the Military
The chief of strategy and communications at the Department of Defense Joint AI Center, Greg Allen, says artificial intelligence offers exciting capabilities for the military’s future.
COGNITIVE TIMES: What was the joint artificial intelligence center set up to do?
GREG ALLEN: We are a very new organization in the Department of Defense. We are the focal point of the DOD AI strategy, and that’s actually spelled out in the strategy itself, that this organization is meant to lead the implementation of it. And our mission is to accelerate the adoption of artificial intelligence into the Department of Defense.
CT: So how many AI initiatives right now does JAIC oversee, coordinate, work on, and then what does that future pipeline look like?
GA: Our two national mission initiatives that are underway right now are Humanitarian Assistance and Disaster Relief and Predictive Maintenance. And these are some really exciting programs. Just to give you a flavor of what we’re up to, in humanitarian assistance and disaster relief, we are running a program to develop a suite of AI models that can automatically analyze the footage coming off drone video feeds or satellite imagery to analyze what’s coming off those video feeds. In the event of a wildfire disaster, the drone flies over the wildfire area, it automatically interprets the imagery based on the AI models’ analysis of that imagery to determine, OK, this is the area with active wildfire. Or this is the area where the wildfire has either burned out or has not yet touched that region yet. And then that data, rather than coming out in a series of, you know, acetate maps that are physically distributed or given updates through radio, it’s actually turned into a map file that is pushed out to a mobile phone operating system to our National Guard partners in the California National Guard. That system is really exciting. There’s a similar one to analyze flood imagery, mostly airborne or satellite imagery, and to determine these are the areas of a region that had been flooded, these are the roads that are impassable in that region. So if you’re trying to plan disaster response activities, you have much more up-to-date information, and that information is available much more rapidly. It’s a really exciting project.
In predictive maintenance, one of the areas that we’re working on right now is engine glass over in the H60 helicopter. This is a problem when you ingest sand into the engine of these helicopters and when it heats up it melts. But when it cools down, it becomes this solid glass. And if there’s solid glass in your engine, you’re in for some trouble. And it’s been historically quite difficult to know, you know, when you might be at risk of incurring that problem. And it turns out that through a combination of analyzing maintenance records and analyzing onboard vehicle sensor data, you can better predict when these types of events are going to happen. So you can move more of your costly and time-intensive, unscheduled maintenance events into your scheduled maintenance events, which are easy to account for in a budget sense and easier to account for in a scheduling sense. So that really delivers a direct impact in terms of affordability and readiness.
CT: How do we compare, when you look around the world to other militaries, other governments out there, how do we compare with regard to AI readiness?
GA: I would point you to the national defense strategy which came out in 2018, and it really reflects some of the best thinking of our experts in this area, in the Department of Defense. And it recognizes that we are in a world of renewed great-power competition. And that the sort of unattainable lead that we used to enjoy over our adversaries and potential adversaries is shrinking, and that has been happening over time. And that’s certainly not a situation that we enjoy or are comfortable with. And so the national defense strategy says that we will pursue AI and autonomy technologies to preserve and extend our competitive military advantages.
I think that’s exactly what we’re trying to do at the Joint AI Center, is to lead in AI technology and lead in the adoption of AI technology to really bring a whole host of benefits to the Department of Defense. Whether that’s sort of back-operations activities and business operations or logistics—which is very familiar in commercial industry or stuff that is, you know, very unique to the domain of national security, such as intelligence, surveillance and reconnaissance and sort of, more tip-of-the-spear-type of warfighting activities. We really believe that AI technology as a general-purpose technology has capabilities that are relevant to all of those applications. And we’re very excited to get some of this stuff into the hands of more and more end users.
CT: What do you think that the DOD does best right now when it comes to the development of advanced technologies, specifically around artificial intelligence? And then what do you think we can and should do better?
GA: Well, I would say the folks at DARPA continue to do a great job. I mean, when it comes to identifying really breakthrough technologies, a willingness to accept risk, a willingness to invest, that all continues to go very well. And there’s a lot of exciting work at DARPA’s third-wave AI initiative, including some stuff that is of interest to the JAIC. But where we have historically had a lot of challenge is not so much with driving innovation but adopting innovation, and ensuring that technologies, whether they come from commercial partners or whether they come from DOD research institutions, ensuring that those capabilities actually get into the hands of end-user communities in a timely fashion. And I think the sort of challenge of technology adoption, and especially AI technology adoption, is really one of the core reasons for the DOD AI strategy and for the formation of the JAIC as an organization.
CT: What do you love most about what you do at JAIC? What is it that you love most about your work and the mission of the organization?
GA: The most exciting thing is delivering capabilities for our end users and seeing the excitement that they have to test our capabilities on an interim basis and how excited they are to actually use the stuff that we’re working on. I mean, we exist to deliver mission impact and talking to our end-user communities and hearing about their feedback and the systems that we’re developing. That’s tremendously exciting. Especially when it’s such an advanced technology. I will say we also have some very brilliant technologists at the Joint AI Center and watching them work is a privilege.