Role of imagery in support of OSINT - Part two

Media Thumbnail
  • 0.5
  • 1
  • 1.25
  • 1.5
  • 1.75
  • 2
This is a podcast episode titled, Role of imagery in support of OSINT - Part two. The summary for this episode is: <p>In the second part of this podcast Robert Cardillo, President, Cardillo Group and previous Director of the National Geospatial Intelligence Agency and ex-deputy Director of the DIA, joins Harry Kemsley and Sean Corbett to continue the discussion of the importance of geospatial intelligence to enhance our use and understanding of OSINT in a classified environment and the use of AI. </p>
Role of Imagery In Support of OSINT - Part two
03:58 MIN

Voiceover: 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: Hello and welcome back. For those of you that listened to part one of this podcast, you'll know that you're about to listen to the second part of our podcast discussing the role of imagery in support of open- source intelligence with Sean Corbet and our excellent guest, Robert Cardillo. Thank you for listening. Yeah, I have to say, as you talked about that assurance and the reliability of things being delivered, it brought to mind times in my past when as an airman I was being called by the supported land commander saying, " Where are my damn aircraft?" You were told to get some aircraft to my, et cetera, et cetera. That does come back to mind, but let's not go down that channel. Let's pivot slightly from the, we've talked about why there must be limits. There has to be some sort of boundary between the commercial and the non- commercial sector. Okay, we get that. Let's turn that around now. Now you are talking to an audience of practitioners in the classified environment, in the geospatial classified environment, and you are seeing that they should or could be engaging more than they are with open source that's available to them. How do you begin to help them understand, A, what's available, and B, why they should get involved with the open source environment. Robert? Because I think we've said in this conversation that in some areas there's no place for it can't be there, it's not going to be reliable, et cetera and other places. We've said it absolutely is, and we've also, I think, indicated that there's probably still scope for more open source in the classified environment. How do you convince them of that? If you are going to convince them, what are you trying to convince them? Let's focus it again on the geospatial.

Robert Cardillo: Yeah, Sean and I both, again, grew up in the operative term of our titles was interpretation. Right? We're hiring you to interpret or exploit information out of an image. Right? It's stuck there in the image. We want you to pull it out and put a piece of paper in a cable, in computer so that people can, to me, this leads to my presentation to the new class of geospatial analysts. I need you to be comfortable and confident that the machine can answer three questions now, very, very well. What, where, and when? By the way, the fact that the machine can do what, where and when is not a diminishment of your job. Because, hang on, I've got two tougher questions for you. We're saving the hard ones for you. You need to find a way, and again, being comfortable doesn't mean you're not challenging and you're not questioning and you're not testing all of those computer vision tools to make sure they are 90 whatever percent accurate. You do understand what false positives and false negatives mean on those returns. But as you can reach that comfort, then I need you to spend your gray matter on why and what's next. Those two questions, now, who knows with ChatGPT and ImageGPT where we'll be in even three months or three years, maybe they'll get there. But right now, it's very, very difficult for machine to do why and what's next. Can it help you? Could it be an intern like or an apprentice like service? That would be my other change. I wish I had thought of this when I was in charge because unfortunately I was rather aggressive with my introduction and disruption and too many of the analysts heard it as us or them, right? Robert's picking the machine over me, he's going to replace my job. I'm a bank teller. He wants ATMs. The whole thing is just going to be I'm out of work. It took me too long to get to what I just described. Again, this is I guess the beauty and the curse of age. You think of these things afterwards, but I do think it's important to have them, the current workforce, get quite comfortable with that machine provision. Because just as Sean described, one of the most interesting things on an image is what isn't there. What isn't there tells you a lot about intentions. Right? A bunch of armored vehicle without the POL trucks aren't as interesting. Right? A brigade or a core that doesn't have their medical unit close by is also a different entity. Right? Now, again, I'm not saying that our friends at OpenAI won't figure out how to do negative absent information contributing to some sort of AI, but right now that's up here in our gray matter. I would want to excite them and I would finish with, by the way, why and what's next are so much more interesting. By the way, much harder. You should be excited and you should be anxious. Now, I'd want you to be positively anxious like, oh boy, let's lean into this. Not, oh my goodness, how am I ever going to figure out why Putin has moved that much armor to that location? What's he going to do? Let's face it, that is at the heart of many of our challenges today to get inside of one man's head. But I do think it's so exciting that the fact that you can at least attempt to tackle that question to me is quite exciting.

Harry Kemsley: The time we've taken in the past, Sean, to get to the point, we actually talked about doing the judgment piece, the actual analysis of the why and the what next actually has been a tyranny, hasn't it, of the lack of abilities to collect and collate and process, which is I guess, Robert, the point you're making in terms of the machine being able to do the collect, collate, and to some extent process to allow you to get onto the really important questions. Sean, I'm going to ask you the same question. Given your background, what do we need to do... And you can't use the same answer as Robert, by the way that's now taken off the sheet. What is it we need to do to get the analysts to understand what could be done in the geospatial environment, what they could be dragging out of the commercial or the open source environment anyway to help them? How do we do that?

Sean Corbet: That's so unfair because Robert nailed it. But it is a continued education. Now, I think in some ways, and I talk about generations a lot, the next generation or the one that's coming through now are probably better equipped to do so because they are inquisitive, but they have got much less attention span. Whereas Robert and I would've been really happy identifying the medical unit and counting tanks, all the rest of it, that's not going to do it for them at all. Getting them to that place where all that inaudible stuff is done, now you've got the fun bit of doing the what if, with whatever source that happens to be, how you do that is inculcating the right culture early on, and unfortunately, you've heard me talk about this before and it was something I really struggled with quite heavily within DIA is that people tend to promote in their likeness. If you toe the line and you do the policies correctly and you follow process exactly, you tend to do well. If you are that maverick that goes, " No, no, no, there's a far better way of doing this and it, okay, it might be a little bit of risk, but look at the answer." Those people tend to be looked at fairly quizzically. They tend to be impatient and therefore move on. But somehow you've got to recognize them and reward them and fast track them onto, they are the people that people then want to follow. In a big institution, government institution, any government institution, that is a really difficult thing to do. You've got to do it by every means you can, whether it's bribery, blackmail-

Harry Kemsley: Coercion.

Sean Corbet: Coercion, flattery, you name it, I tried all of those and the most, the one that worked more than any of the others was allowing people to use my car parking slot.

Harry Kemsley: Okay, car parking's always good. I think what you've just said, Sean, is the digital native, the technology natives that we're seeing as the younger generations who have no attention span, I think is what you said. We need to get them to a place where they can actually move quickly to the things that really are interesting, which is what Robert's talked about. I think we've said that the open source and commercial environment might be able to help them get there. I think that's what we've said, and technology certainly being a big piece of that. Let me then move the three of us onto the next part and maybe the last part of this conversation for now. We've seen over the last 40 years, Robert, you've described it very well for us earlier, the trends in capability and how that's driven the foundational flaw of what I start with in my analysis rising all the time around us. Where does that stop or is it going to stop? What are the big trends right now that you are seeing coming through that really excite you in the geospatial environment? Probably because some of that will be in the technical means that are classified. Let's stay out of that realm. Let's stay firmly in the commercial realm. What are you seeing coming through that for you is an exciting part of what's going on in the geospatial world and how do we get it to the next level going into government?

Robert Cardillo: Yes, this'll be fun. A couple things. One, I am quite comfortable with what I believe is the reality of ubiquitous sensing. I appreciate that some in the audience will shudder and go, " I don't want to live in a world that's ubiquitously sensed. I have enough anxiety with government tracking, monitoring-"

Harry Kemsley: Facial recognition and so on and so forth.

Robert Cardillo: Yeah, sure. Right. We need more anonymity and I'm not disagreeing with that, but well, I understand that sense, that negative anxiety that could be coming with it. I guess my counsel would be we, and that's a big we, because this is a social compact discussion about how we're governed, what limits we set on the government. The US Congress is in the midst of a current debate about the accesses of our Federal Bureau of Investigation with respect to said surveillance. There are strong views on both sides, right? One side says it's a necessity to keep us safe. The other side is strongly or more strongly believes it's a fundamental infringement on my individual liberty. By the way, I think it's a beautiful tension within our systems, the power of individual liberty and the necessity of collective security. Okay, figure that out. Those are two things we both want, but what you just asked Harry is about where's that line going? Well, I do think there is coming, and look, I know we're on a podcast, but I'm going to hold up my iPhone and say, we've really already agreed to ubiquitous sensing by carrying these things with us. We have decided individually to make a trade. I am going to trade the fact that a government or a human or a bad actor can now follow me, know where I'm at or where I'm not at, when I'm places, even what I'm doing and who I might be meeting with. But I'll trade that for the convenience that I get back. I've made a deal in this case with Apple, other people that have made deals with Google and Huawei or whomever, fine. Let me get back closer to your question, Harry. I think that's just going to continue. Again, I work with a company that is scanning the earth once a day. Planet Labs, the second entity that will be able to do that as Chinese, inaudible. They have about half their orbit up now in next year or two. They will have global access on a daily basis. By the way, commercial unclassified, you can go buy it or rent it or lease it as you wish. That's not going to stop. People know about the Starlink and Kyper, the thousands of satellites. Anyway, that's going to continue to happen. Where I am not sure, and this I think goes to what Sean was saying about what era, what generation you're from. I unfortunately think the government is still acquiring commercial in an additive way. Okay, here's what I have. I own this. I operate that. I'm going to go buy an increment of this and add it to the end of my equation. As I've said earlier, I think that needs to be fundamentally flipped. I also am well aware of how unusual and how difficult it is to do a fundamental flip. The reason is, and it isn't just the biases that we've talked about, and at least in the US system, it's the tyranny. Does Robert want to say tyranny? He just did. The tyranny of our acquisition process, our programs of record, I am a huge fan of Robert McNamara and what he built in the sixties with our future year defense program. We call it inaudible. Well, we built that program because we were fighting a machine called the Soviet Union. We built a bigger, better machine and our machine won. Right? We just out Sovieted the Soviets. I'm sure I'll be okay with that quote too. Let's face it, we went through a non- peer threat, we're still going through. The non- state actor, but the pacing threat for all of us now is China. Yes, it's a big system, but it is not the Soviet system. I worry that we've kept the model that we developed to beat the Soviets and we're trying to apply it against the Chinese, and I fear it will not work. By the way, what I just said intellectually, I'm sure my colleagues in government agree with it. The problem is there's so much momentum in that machine that I mentioned. By the way, the machine is fueled by companies, some of whom I work for now, by Congress, okay, by lobbyists, by the natural momentum of the Pentagon's requirements process. Again, there's nothing wrong with anything I just said. No one out there is saying, " Oh, I wonder how we could figure out a way to lose the China." It's not what they're trying to do. It's just that sometimes we can have these conversations on podcasts like yours, but then bringing them back into the building is enormously difficult because there's so much counterweight to everything I just said. Apologize for bringing depressing the conversation there, but look, I don't think we can deal with it unless we talk about it and tee it up in that way.

Harry Kemsley: Well, I'm going to have to ask then. So I don't disagree with what you've said. For a period of my life, I choose to forget, most of the time I worked in capability development in Whitewall in London. I know the sort of inertia and problems you are alluding to all too well. But in a few words, how do you shatter that? How do you break that cycle and get to a place where we are going to be able to keep pace, if not get ahead of the anticipated threats that we're seeing growing in front of us? How do we reverse the trend you've outlined, Robert?

Robert Cardillo: I'm not optimistic. A friend Chris Brose wrote a book about five years ago and called The Kill Chain. I recommend it because Chris covers well, and he had tons of experience on The Hill about that inertia us kind of doing the same thing over again, hoping something else comes out the other end, which is another definition for insanity. But here's where I think, and this is why I'm not optimistic, as good as Chris wrote in that book in describing the challenge of requirements based processing and long- term development and acquisitions schedules, he barely touched on our Congress. I don't want to lay all the blame there, but I'm going to lay a good bit of it there. There's a challenge, at least in the US system that look, it's a balance. We try to have three branches of government that can check one another. That's not an accident. Unfortunately, the check that Congress has is on money and timelines. Look, we may not have a budget at all in FY24, which starts in about two and a half months. Meaning Congress will push through what's called a continuing resolution, just, oh, you'll have the money you had last year, but no new starts. Right? No new builds because we haven't authorized them. China couldn't do a better job at holding us back. Again, do I think Congress is up there trying to help China? Of course not. But because they want to hold onto their power, I have a check over the executive branch, I'm not giving that up because that's how our system works. You can't do something new until I give you permission. That's why I'm not optimistic Harry. I don't think there's going to be any real change to that system, unfortunately, until we have a huge wake up call and I worry about it. Whether you call Pearl Harbor, 9/ 11, I worry that... We will survive the inaudible. We will still exist, but could China make such a move so quickly that our lesson is just too late?

Harry Kemsley: Yeah, we've learned a lesson, but we can't implement it. We're not here to implement it. Sean, it's your task. Now, with that same question, looking at it from a non- US perspective, looking into other parts of our community, let's talk about NATO. You've worked in NATO, you've worked in the UK. Do you see the same sort of inertia? If so, how do we turn that around outside the us? I mean, Robert's talked about Congress as being part of the problem. I'm sure there are other parts that he could talk about, but he's zeroed on Congress for now. What about outside of the US? Is there the same sort of inertia? If so, how do we break that problem?

Sean Corbet: The answer to that is firmly. Yes. I'm even more gloomy than Robert is actually, which is I tend to be, anyway. Again, it's a point well- made that everyone's aware of this and everybody thinks, yeah, we need to do something about it. All that happens is new people come in, they try new processes, which is still incremental, which is still process driven, which is still authoritative. Again, I agree, and we've mentioned this on previous podcasts, that I really hope it doesn't need a significant strategic shock to get it to because it might be too late, because each progressive strategic shock we have is worse. It might be existential, which I don't use that word very often. How do we do something about it? By doing this sort of thing at a very, very low level to try and I don't know who listens to this in the movers and shakers within the MOD, for example, if anybody. If you listen to the current head of strategic command, as I have several times when he's been briefing, he is very upfront saying, " Look, we need to collaborate more with the industry. We need to work better with industry." Which is only one small element of it. I think one of the reasons he's saying that is now it's the industry that are the innovators, the people that are agile enough to move. The question on that though is are they going to get the investment they need to actually capitalize on that and keep on doing that. There is something there for the commercial sector, but that doesn't answer everything either. Now we could start going down the rabbit holes of, okay, do we need dedicated procurement people? Because we've all been in a position of capability development where I know you were much better than I was in there, but you're an amateur trying to develop something that is incredibly complex, and that's just the process, let alone what you're actually trying to deliver. Do we need specialists who stay in there all the time? You can say, " Well, yeah, we do have them, the civil servants." But you need people at the decision level to do that. Is that, and I noticed actually from the air power conference's just read out, is that there's initiative to try and get people to go from industry to the military back to the industry, I think a zigzag approach or something to try and get that agility, whether that will break the mold or not, I don't know. It's an impossible question to answer.

Harry Kemsley: All right, well I do think you've touched on something though there, and at the risk of sounding like I'm trying to paint you both into a corner, I think what we've said is commercial is moving at a pace that government isn't. We've said that governments need to move more quickly, and we've just said that we probably need more commercial in government, but there are limits. I think that's what we've said in terms of how the open source environment might help government. There is a podcast waiting to have that conversation though but I'm going to start to draw stumps on this conversation from fear that the listener's going to run out of coffee listening to it. Let me do this. In a second, Robert, I'm going to ask you to give me your one takeaway from this conversation you'd want the listener to actually walk away from this conversation within their heads. What's your one takeaway? We're focusing here on geospatial and the open source environment. Sean, I'll come to you to do the same. Then given that I'm moderating this, I'll take the last chance. Robert, where do you want to go in terms of your one takeaway for the audience?

Robert Cardillo: Audience, despite the last 10 minutes, I would hope the audience could leave uplift in this way. There's so much potential power in this sensing, and I'm not going to cast it in a military or even in national security. I just strongly believe that the more transparent our planet is, the better it is for open societies. Even though I absolutely value individual liberty and personal privacy, and so we have to find ways to condition that access and whatnot, broadly speaking the more transparent, the better. Putin does not want a transparent planet. Xi does not want to transparent planet. That is not part of their social compact with their citizens. For all of the challenges that we've described today, I would ask the audience to be broadly embracing of that transparency and help us find the guardrails that are needed in places but don't wish it to go away. That's probably not a good course of action.

Harry Kemsley: The genie is out of the bottle.

Robert Cardillo: Don't hope that the science begins to regress. That's probably not going to happen. I guess I'd leave it with embrace it. Right? Interact with it. There is always my strong belief that that human component will always be exactly what's needed. Yeah, jump in and make it better.

Harry Kemsley: Thank you. Thank you Robert. Sean?

Sean Corbet: As always, I'll follow Robert Cardillo's lead and have a positive approach, which is unlike me as you know. I think back to the incredible increase in capabilities we've had the full spot spectrum, sorry, of open source intelligence, but of course with imagery being my first and lasting love, I'm going to focus on that because I think you are seeing the greatest innovative developments. It's positive. That understanding, that ability to understand the world is so much greater now. There's no question about it despite the disinformation, misinformation, obfuscation as well. But I think the key is to, as much as we can become comfortable with the risks that entails. All the risks that Robert was talking about in terms of everything from the security implications, how assured is it really, we've got to become more comfortable with that risk and to maximize its impact.

Harry Kemsley: I think for me then, the one that I'm going to take away from this conversation is that I have been heard to say on this podcast many times, we need to get more engaged with open source and not just the commercial, but the open source environment. We've just got to get better at it. I think what you've done for me today, Robert, is you've underlined and reminded me, yes, of course, but there has to be limits. There has to be a space in which the national security of the company, the organization, excuse me, that we're working for, the nation we're working for, got there eventually, that national security has to be dependent on things that are totally assured and in our control. There has to be a limit, and I think that's a piece that I'll take away from that. The open source, of course, it's important, of course, but it can't be the thing that I necessarily use for my decision making. It has to be somewhat more bounded for that, for my national security. With that, then let me stop and say, Robert, a huge thank you for your time. I know how busy you are. I can only imagine how busy you really are. There aren't many people in the world that could sit here and have this conversation with us about the geospatial environment that know more about it than you do. You've been at the very beginnings of the recent generations and you I'm sure will be an influence for many years to come. Thank you so much for your time today. A great conversation.

Robert Cardillo: It's been fun. It's a pleasure. And at the risk of complicating future calendars, if and when you all want to pick up the next chapter or next couple of chapters on this conversation or narrow it down on a certain area, please, please, I'd be happy to continue the conversation.

Harry Kemsley: Well standby for incoming on your calendar because I can guarantee that's going to happen. Robert, thank you. Sean, thank you as always for joining me, and to the listener for taking the time to listen to us talk. Thank you again.

Voiceover: 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.


In the second part of this podcast Robert Cardillo, President, Cardillo Group and previous Director of the National Geospatial Intelligence Agency and ex-deputy Director of the DIA, joins Harry Kemsley and Sean Corbett to continue the discussion of the importance of geospatial intelligence to enhance our use and understanding of OSINT in a classified environment and the use of AI.

Today's Host

Guest Thumbnail

Harry Kemsley

|President of Government & National Security, Janes

Today's Guests

Guest Thumbnail

Sean Corbett

|AVM (ret’d) Sean Corbett CB MBE MA, RAF
Guest Thumbnail

Robert Cardillo

|President, Cardillo Group and previous Director of the National Geospatial Intelligence Agency