Role of imagery in support of OSINT - Part one

Speaker 1: Welcome to the World of Intelligence, a podcast for you to discover the latest analysis of global military and security trends within the open- source defense intelligence community. Now onto the episode, with your host Harry Kemsley.

Harry Kemsley: Just before we start this podcast, just a quick explanation that for this episode we're going to split it into two parts. This first part of the role of imagery in support of open- source intelligence will play now, and then we'll invite you back to join us for the second part very shortly. Hello, and welcome to this edition of World of Intelligence at Janes. As usual, Harry Kemsey, your host and my cohost, Sean Corbett. Hello, Sean.

Sean Corbett: Hi.

Harry Kemsley: Sean, like you, I have been watching events around the crisis in Ukraine and seen a huge amount of open- source material, but amongst all that, we've seen a lot of commercial imagery, a lot of geospatial intelligence coming through. So I thought today we might spend a bit of time looking at geospatial intelligence within the open- source environment. Now, if any of the listeners have been anywhere close to geospatial intelligence in the recent years, they will come to know the name Robert Cardillo, and I'm absolutely delighted to introduce Robert. Hello, Robert.

Robert Cardillo: Good morning. Greetings.

Harry Kemsley: Good morning.

Robert Cardillo: How are you?

Harry Kemsley: Fantastic. And thank you so much for taking the time to join us this morning. Robert Cardillo is the president of the Cardillo Group, founded in May 2019, delivering strategic consultation services dedicated to the growth and development of the intelligence profession. Robert is engaged in numerous nonprofit services and holds a range of board and advisory board positions with organizations across the defense and national security spectrum. He's the only person to lead the analytical operations at four different organizations in the US intelligence community. In reverse chronological order, until February 2019 he was the sixth director of the National Geospatial- Intelligence Agency, NGA, leading the transformation of the agency's future value proposition through innovative partnerships with the growing commercial geospatial industry. Prior to that, within the Office of the Director of National Intelligence, ODNI, serving as the first Deputy Director for Intelligence Integration from 2010 to 2014, Robert managed and delivered the President's daily brief to President Obama and Vice President Biden, and was a member of the Deputies Committee of the National Security Council. From 2006 to 2010, Robert served as the deputy director of the Defense Intelligence Agency, DIA, and the deputy director for analysis of the same agency. In 2009, he served as the Acting J2 Intelligence Lead, which was a first for a civilian, in support of the Chairman of the Joint Chiefs of Staff, Admiral Mike Mullen. Robert's distinguished service has been recognized at the very highest levels. In addition to an Honorary Doctorate of Humane letters from St. Louis University, Robert has been awarded the Presidential Rank of Distinguished Executive twice, the Presidential Rank of Meritorious Executive, the Director of National Intelligence Distinguished Service Medal twice, the Secretary of Defense Distinguished Service Medal, also twice, and the Chairman of the Joint Chiefs of Staff Joint Meritorious Civilian Service Award. So Robert, from that bio, it'll be very clear to the listeners that didn't know you, your incredible knowledge and understanding and experience in geospatial intelligence. Perhaps I can get you started by asking how did you get into geospatial intelligence in the first place?

Robert Cardillo: Happy to, and Harry, Sean, it's very good to be with you and with your audience here today. I know we're going to have a little bit of fun and hopefully-

Harry Kemsley: Absolutely.

Robert Cardillo: ...nudge the needle ahead with respect to advancing our overall profession and contributing to decision makers and better outcomes for all of us. I have a short answer of how I got involved in this profession, it is by accident. But of course I've got a longer answer that that's a little bit more interesting. Coming out of university I knew that I wanted to serve somehow. I came from a service- based family with my father in the military and my brother. But the uniform didn't appear to be the right fit for me, so I applied to the Defense Intelligence Agency in the early'80s. So you now dated your guest today. Ronald Reagan was in his first term of office. He was rebuilding US defense, as you'll recall, and I ended up putting my name in at the right time and place because the Defense Intelligence Agency had decided that they were building all of these highly classified imaging satellites, but they didn't have enough humans to exploit the result, the collected imagery of said satellites. So they said, " Oh, well, we might want to go find some more," what they called in those days, " photographic interpreters." And so Robert was in the right place at the right time. I was a warm body. I could put two and two together, and they welcomed me in. I had no background. I didn't study remote sensing or geography in school. So not only was it a bit of an accident, I got lucky and I understand that, but what a ride I began. First of all, Harry, in those days when I was welcomed into the profession, I walked into a dark room, usually in the basement of a building with no windows, and I had a badge around my neck that told people that I had a top secret clearance. And then I would of course go through my training and my education to be able to extract information from imagery, and we can talk more about that as we go on. But once I could create value, this is how the process worked. Another human... Oh, and by the way, I apologize to my British colleagues. The other human usually had to be a US person with a top secret clearance, and they would ask me a top secret question. I would task a top secret satellite. I would get back a top secret image, I would put it on a semi- secret light table. So now I'm looking through a microscope that said" top secret image," and I'm employing my craft. I'm extracting information from that. I would finish the extraction. I would turn to a very antiquated top secret computer. It would wait for it to warm up. I would wait for it to please stop blinking and let me enter my characters into the green screen. I would hit Complete, and there I would have my record. " At this Soviet base on this day, Robert counted this many GAZ trucks and T- 60 tanks," and et cetera, et cetera, whatever the object of the realm was. I say that to you just to frame the audience that yes, that's 40 years ago, but it's only 40 years ago. So two generations ago, everything that we're about to talk about on this podcast was only possible in those basements, in those windowless offices, with everybody wearing a top secret clearance. And so talk about a very closed world. Now, I described that for two reasons. One, as closed as it was, I would argue it was enormously valuable to US and allied security because it was very, very difficult to do what I just described, and enormously expensive. I mean, one, access to space obviously. We had the space race with Sputnik and the Apollo missions and whatnot, and yes, were the Soviets putting cameras in space as well? Yes, they were. But our attempt and interest was to put better cameras and have those... And by the way, those cameras in the early days were literally cameras, which would eject film. I don't know how many of your audience will know what film is, but it's a hard copy record of a light source interacting with a piece of plastic that has the right chemicals on it. I know I just sound 1, 000 years old now. But again, I marvel at what we're able to do, to eject that film at 25,000 miles per hour in space, have the film fall through our atmosphere, have it deploy a parachute, and be caught over the Pacific Ocean by a Air Force plane with a trapeze behind it. I mean, it sounds absolutely James Bond incredible, and I would argue it is. And so we should pay great tribute to those pioneers that created that ability to sense our planet from space. We have come a long way, and I'm now going to fast- forward and turn it over to Sean to bring us up to speed with respect to how we think about and how we define this thing, geospatial intelligence. But I'll finish with, for your audience, it really is all about sensing the world in a way that makes it safer. Whether that's for navigation, whether that's for transportation, whether that's for simple assuredness of my security of where I am and when I am, but obviously, because of the importance of location and having that sense of security, the profession has grown to the term now, geospatial intelligence. Sean.

Harry Kemsley: Sean, geospatial intelligence. Oxford English Dictionary, go.

Sean Corbett: Thanks for that, Robert. It is a trip down memory lane, and our backgrounds are extremely similar. I was

Harry Kemsley: I was going to ask, weren't you also a photographic interpreter at exactly the same time?

Sean Corbett: I joined the military as a photographic interpreter. And it was only much later that, A, we had intelligence plans, but we actually changed to imagery analysts. And then of course, finally the geospatial intelligence piece came in. But I do very much remember going into the dark rooms with the photographers and splooshing, as we used to call it, say, " Right, can you just enlarge that particular tank," whatever. And whilst I'm even older than all the rest of you, it's not that long ago actually, but just bring us up to speed on terms of the terminology. So we've gone from photograph interpretation to imagery analysis, and now the term is geospatial intelligence. And if you want a formal definition, it's effectively the exploitation analysis of imagery and geospatial information to do the usual describe, analyze, assess, and visually depict physical features and geographically referenced activities on the Earth. Everything is somewhere. And so being able to situate it, if you like, in terms of where it is and what that means, is the way forward. So there's three elements to do, and the imagery itself, and we'll probably talk about the vast different types of imagery now. There's the imagery intelligence, the interpretation of what you're seeing on the image, and then the overlaying of the geospatial information is bringing them all together, is what makes GEOINT.

Harry Kemsley: Right. Well, let me get us started then. You used the word" closed" quite a few times in your introduction, Robert, and now of course we're talking about open- source information. And I think one of the big things that's changed in the last 40 years is how much of what used to be the exclusive domain of the government, the exquisite capabilities of government, were the only ones that really had imagery, certainly from outside Earth looking in. Today, of course, that's less true. We can come back later to talk about whether that's actually going to be sustainable, but how much of what you used to do, that was entirely closed and top secret, is now open? Is it all of it or just still a subset of it from your estimation, Robert?

Robert Cardillo: It's a good question, and I'm going to take a risk and put some numbers to my answer, some percentages, but please know, these are just gut reactions. I was an analyst from like'83 to '93, so up and through the collapse of the Soviet Union, into the early era of a truly unipolar world of the'90s, and the beginning of more of non- traditional analytic efforts, special operations, counter- terrorism. Of course, terrorism in Robert's day was ship jacking and hijacking, hostage- taking, but still same threat to our society. I would say easily in my early days, especially when I was doing Soviet analysis, 100% was coming from government sources, classified sources. Now, I just decided to take a footnote there on my own. Of course, I used maps and charts to baseline and contextualize my analysis, and so those were unclassified and sharable and those were derived from open sources as well. So maybe I would go to one, two or 3% there, instead of just 100. But I will tell you, my mentality was, if I'm going to provide value, if I'm going to give advantage to the person that's going to read my report, it's going to come from that classified source. I had no inkling that I was going to get anything of value that wasn't classified. I ran NGA as director from'14 to'19, so that's more recent, last 10 years. And again, by the way, Harry, we can go back, I've just jumped over 25 years, but-

Harry Kemsley: It's okay. We're on a timeline inaudible.

Robert Cardillo: ... I'mjust giving you my sense of where the analytic frame of mind was when I was director. It's a big agency with thousands of analysts, and I'm now going to generalize so this is a homogenization of a lot of efforts. But I would say that there were some parts of the analytic organization, certainly the team that would work humanitarian issues and disaster relief and recovery and climate monitoring, preponderance of their work was open. Whether it was US commercial imagery, whether it was national imagery coming from places like NASA or from Europe's Sentinel missions. And let's face it, we were introducing more and more airborne collection throughout the time period. And generally speaking, the old role was, it was in space, it was top secret. If it was a US or an Allied plane, military plane in altitude, air breathing, secret or collateral. But even then it was much easier to make it releasable and unclassified. So easily, those types of accounts, the majority could be open- source. But if I went back and if I visited my China shop at NGA and they're tracking Chinese mobile missile deployments and exercises and whatnot, they would be using now our partners at Maxar and now BlackSky and Planet as baseline, the way I would use it as baseline as maps, but it wasn't differential. And we should come back and talk about why that is. There's still a heavy bias, and that's the right word, it's a bias, that I think needs to be disrupted. And again, we can talk more about that, for not just government, but classified. Unfortunately, in the minds of my colleagues, and some of my colleagues of my class are still exploiting information today, there is an unfortunate belief, and I say unfortunate, it's reality, but it's carried forward for too long. And this belief is this, the higher the classification, the more the value. " My source is top secret. Your source is only secret, so mine's better than yours." And yes, it sounds silly, it is silly, but it's there, and it is a problem to deal with. And so I think one of the challenges that we've had to have commercial fully, and I said commercial open fully embraced, there is unfortunately that lingering bias. How could anything unclassified be useful to me?

Harry Kemsley: And Sean and I have certainly talked about that before, haven't we, Sean, where we talked about the lack of engagement with the open- source environment. And by the way, Robert, we talk about open- source, we capture in there the publicly available and the commercially available. It is what's out there outside of government, if you will. And we've certainly talked about that, Sean, in the past, haven't we, about the lack of engagement. I don't think it's as bad as it would've been 30 years ago, or maybe even 20 years ago. It's getting better, but there is still not a full engagement, is there, Sean, in terms of what's available and what could be used from the open- source environment?

Sean Corbett: No, we're not there yet. And that's for several reasons, one of the major ones is, which again we've talked about, for cultural of course. But in answer to that original question, and this will come to answer that as well, in terms of what you can use in the open- source, I'd go back to the way that Robert characterize it and go, it's in comparative terms. That's stuff that we were both looking at in the deep, dark recesses of top secret buildings many years ago in the'80s. At highly classified levels that nobody could see, you fast- forward today and some of the capabilities are astonishingly good in comparative terms, and that's the way I would do it. So not comparing like with like today, but some of the 30- centimeter plus color images, the development is great. And of course it's all about what you're going to use it for. We've got so much now in terms of the data, the key, and we probably should definitely come onto this, is that knowing where to look, knowing why you're looking there, knowing what you're looking at, and we might definitely go about this one because not everybody can be an imaging analyst in my humble opinion. I would say that, wouldn't I? And then what is the problem set that you are trying to resolve? So that answers the piece about why the interaction isn't there. And I will use the word" paranoia," which I think is a little bit strong, but a paranoia, saying that, "We do not want to reveal our intelligence gaps, therefore we are not going to tell you in the open- source what we're interested in and why we're interested in it." Now you can actually obfuscate and you can hide that to be fair, and I really do think you can, but there is a reality there. Of course, people don't want to do that, particularly in the imagery side where all these satellites going around the world are being monitored very closely. So those people who we might be interested know exactly when a certain pass is going to be. And I'm talking about commercial satellite imagery, of course here.

Harry Kemsley: Let me just pick up on something you said as well, Robert. You talked about that foundational contextual understanding you got very early on from maps, that would give you some sort of context. And then you talked about how commercial imagery has become that foundation. Do you see a time when what's foundational actually becomes more actionable? It's much more directly going to be part of your product at the end of it. And just to pick up on the point Sean made there about the resolution that's available now and the incredible image you can get from commercial sources, probably being better than some of the stuff you saw at the very early stages of your career from government and then exquisite sources. Do you see that trend continuing to the point where actually you don't need government satellites to do what commercial satellites are now doing?

Robert Cardillo: Your last sentence set off a chill, Harry, because-

Harry Kemsley: That's why I asked it.

Robert Cardillo: Well, look, when I came back to NGA, I grew up there, went away to Defense Intelligence Agency, went down to the Joint Staff, did my White House assignment for President Obama, and really came back to an agency that I had been away from for 10 years. And I was quite worried as a ... Well, I was excited obviously to be able to be the director of the agency where I was once the most junior employee, so that's quite a gift to be able to have both ends of those experiences. But I also, and Harry, trust me, I'm going to circle back to your question, but I was worried about what I thought was a false and somewhat dangerous comfort at NGA into 2014. And it went this way, it was basically... And you can think of a lot of industries that have had this challenge. We were successful yesterday, that was the mentality. We applied the tradecraft. We had exquisite access to space. We had the best sensing technology in the world, and we won, or we overcame or we provided advantage. And we'll just do that tomorrow. And if we need to pedal faster or try harder, we'll do that. But that's what I thought was the risk, because everything you just said, Harry, about what I call the growing transparency of our planet, meaning the sensing is almost continuous. Which by the way, Sean, can help us on this whole, " What do you mean you're looking over here?" Well, everyone's looking everywhere, so that's a noisy environment and it's easier to cover. " Oh, I'm really interested in that spot there, but I don't have to let anyone know because there's so much sensing going on." So I tried, and I'm sure the jury's out a bit at NGA and maybe they'll write history books someday about how well I did, I tried to disrupt that comfort. Now I intended to do it constructively so the disruption would lead to advancement. And look, it's a 10, 000 person organization. Did some people run to the corner and hide and wait until I had my ceremony, had my turnover? Of course they did. And are they still there? Of course they are. But by the way, I don't-

Harry Kemsley: inaudible?

Robert Cardillo: No, no, no, no. And I was just going to say, I actually don't think that they're bad people or somehow working against our security interest, they just saw me as a risk because that openness was going to water down, was going to diminish the special value that they had become comfortable with. And I thought just the opposite. I said, " Look, this isn't about elimination, it's about elevation." It's about taking all of this now open sensing... And by the way, the term I use is unprotected data, because there's protected data in many different forms. But if you could just access that unprotected data, the base foundation that I started my work on, that 1: 200000 scale map of the Transcaucasus military district, is enormously enriched now. You're starting at a much higher information point. It is a little scary to these analysts because, " Wait a minute, if everybody knows that, why do you need me?"

Harry Kemsley: "Where's my value?" Right. "Where's my value?"

Robert Cardillo: That's right. But this goes to your question about do we still need government sensing? My answer is adamantly yes, it should be I'm going to say smaller. By the way, smaller doesn't mean less expensive. Our governmental assets continue to be quite expensive.

Harry Kemsley: Indeed.

Robert Cardillo: And it doesn't mean smaller is less important. In many ways, let's say we go to a world where 95% of the source of an NGA analyst is open. Okay, great, all right. They're really at a very high point of information and awareness about their target or their topic. That 5% that comes from an exquisite, classified, compartmented access that is really far edge of science or of space access, great. But that's the problem. When I did my analysis, again, a million years ago, I would finish with my classified work and sprinkle commercial on top. Literally just sprinkle it on top and go, " Oh, look at that. Look at me. I've done my commercial bit for the day." That has to be so fundamentally flipped. And again, I worry because I don't believe we've changed the tradecraft development and education courses this way. And I was king four years ago, so you could ask me next, " Why didn't you do this?" But I think we should take new entrants into... And boy, I think you could almost keep them outside of the SCIF, so the classified environment, for a good six months. And just have them completely absorb what that open has and then bring them in much later. Because I want them to think about their tradecraft that way, " How can I share openly and what unique can I apply at top of that?"

Harry Kemsley: Yeah. We've certainly spoken, Sean, haven't we, recent times about maybe Ukraine has made us start to realize that the open- source environment has come of age. We've certainly talked about the shareability aspect of open- source being a real advantage in certain respects, when you're working with non- traditional allies. But let's just go to this point though, Sean, from your perspective, because you've got as much experience as anybody in this conversation, certainly more than me in geospatial, where the room for the value- add for the analyst is becoming compressed. I mean, Robert, you just talked about 5%. Is that compression going to continue? Are we going to get to stage where the analysts who are working in the classified environment, that bit they're adding on top is getting smaller and smaller. Is there actually a chance, the second part of the question, they get overtaken by what's available in open- source. Unless you start to feel as though they're falling behind what they could get. The exquisite just doesn't give them any advantage. Do you see that as a trend that could end up?

Sean Corbett: I mean, that's a big question. Again, probably another one for a whole podcast, but I would address it in terms of context about the whole of the intelligence integration piece. Because I do feel that there is a place still for the specialist imagery analyst. I really, really believe that, and I've seen that myself. Everyone's familiar, well, certain people in the West, are familiar with the certain viewpoint of the world taken from the air, if you like, or satellites. And so it is not difficult to say that's a tank or something else. But taking the context of the fact is there are these vehicles and there's a couple of signature equipment here. They are in this position, which means that they may be about to do something. The other layers of the intelligence, you still need a specialist. Now, whether that's an open- source specialist or a classified specialist, and it goes back to Robert was saying, that 10%, which I think is absolutely firmly the space that the intelligence community should be looking at, that's where you can add... You can take all the foundational pieces, but that's where you say, " Right, this is the absolute key piece of information that we can only get from classified means." That's what adds a value there.

Harry Kemsley: Okay. Robert, in terms of the world of intelligence that you've come from, let's go back to the vault. Inside the SCIF, the world in there has become... The floor, the foundation has got higher, I've got less to go and find out for myself. I'm now adding that five to 10% on top, to sprinkle on the commercial, not the commercial, the exquisite, to make it just what I need it to be. How much of that last five to 10% could be brought in from the commercial sector, but we choose not to? Are we excluding it because we really can't afford to bring them in, it's just not safe to do so, or is it just not available yet? I'm talking specifically the geospatial world. Are we excluding stuff or it is just not available out there, what we need?

Robert Cardillo: Yeah, no, I understand, Harry. I first need to clarify, it isn't 90% commercial inside the SCIF now. It just isn't. It's like I said, parts of the building, yes, maybe, depending on... The account parts are probably 50/50.

Harry Kemsley: inaudible.

Robert Cardillo: I would say if you went back into the China shop, I think it's still predominantly national technical means, which means government- owned, government- operated, government classified. And if I could add one thing to Sean's good comment about the tradecraft piece. Look, I just came out of the Oak Ridge, did a Trillion Pixel Challenge, and it's just another computer vision. How quickly can we process information or these supercomputers and whatnot. And again, interestingly enough, I work with a company that collects the imagery, called Planet, and I'm now working with a small AI startup called Synthetic. And they did some interesting work on tracking that Chinese balloon back through to Hainan, over time, through the archive. And when they found out about the challenge, they said, " Well, we used 18 trillion pixels to do that. Why is the government simply doing 1 trillion?" I just love the story that our government is trying hard to be hip and happening, and universities and students are going, " Boy, these guys are old." Now I need to make some calls and apologize to my good friends at Oak Ridge. So back to your point, I think there's something else we need to bring up here. I talked about the inhibitor, about the bias of classification. There is also a bias, and it's not... Again, just like classification isn't crazy, this one isn't either, and this one is control. Here you're a government official, you're responsible for a national security mission. Lives are at stake if you do or don't do your job. And I come into you and I have some commercial offering for you, and you're going to go, " Okay, I'm going to sign a contract with this company and that contract's going to say that they're going to send me these things. And they're going to assure me that none of their pixels or their ones and zeros have been manipulated or in any way affected." There's going to be clauses in that contract that talk about payments and about assuredness and about cybersecurity. " It sounds risky to me, Robert. I really can't put my mission at that kind of risk." Okay, so that isn't a crazy thought. You obviously do want assuredness, you do want... And there are many... Many. There's a good portion of senior leaders in government who will at least resist, if not reject, commercial for just that reason. " I can't be confident, one, that they'll be there when I need them, that they'll be responsive to my,'I have to have this in 10 seconds, not 10 minutes.' And look, they live out in the wild. I mean, I feel uncomfortable enough with my cybersecurity here in this SCIF. And you're going to tell me that company in downtown San Francisco, that is in the open, is going to not be affected by an adversary as capable of China or Russia. Are you kidding me?" So we need to bring that into the equation, Harry, as well. By the way, I think it's eminently addressable. Obviously, nothing in life is assured. And by the way, that goes for government and commercial. But what I like to take my friends who bring it up, I said, " Look, I get it. And you should absolutely build in every protection you can into that exchange or that contract." But don't kid yourself. A one and a zero could be a zero and a one in a millisecond, inside your SCIF or out. And so we're going to have to figure out how to close this circle in a way that we can confidently move forward.

Harry Kemsley: Okay, we'll take a short break there. That's the end of part one on the podcast discussing the role of imagery in support of open- source intelligence. Please do join us for part two, where Robert, Sean and I, will continue to discuss the geospatial collaboration between private and public sectors. Thank you for listening.

Speaker 1: Thanks for joining us this week on the World of Intelligence. Make sure to visit our website, janes. com/ podcast, where you can subscribe to the show on Apple Podcasts, Spotify, or Google Podcasts, so you'll never miss an episode.