How emerging technologies drive the development of open-source intelligence

Media Thumbnail
  • 0.5
  • 1
  • 1.25
  • 1.5
  • 1.75
  • 2
This is a podcast episode titled, How emerging technologies drive the development of open-source intelligence. The summary for this episode is: <p>In this episode of the Janes podcast we look at the application of emerging technologies to enhance the value and to drive the development of open-source intelligence or OSINT. For example, we look at the use of artificial intelligence and using algorithms to make sense of enormous data sets.&nbsp;</p><p>Michael Horowitz is the director of Perry House and the Richard Perry professor at the University of Pennsylvania. He's an acclaimed author of books and peer reviewed articles, often based on his research interests in the intersection of emerging technologies, such as artificial intelligence and robotics, with global politics and military innovation.&nbsp;</p>

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: Okay, good day. My name is Harry Kemsley from Janes. I'm the president of national security and government for Janes. Today, we're going to be looking at the application of emerging technologies to enhance the value and to drive the development of open- source intelligence or OSINT. My co- conspirator today as usual, is Air Vice Marshall retired, Sean Corbett. Hello, Sean.

Sean Corbett: Hi, Harry. Good to be here again.

Harry Kemsley: Good to see you. And we're delighted to introduce our guest today, Michael Horowitz. Delighted to have you with us Michael, thank you for joining.

Michael Horowitz: Thank you for having me.

Harry Kemsley: Michael is the director of Perry House and the Richard Perry professor at the University of Pennsylvania. He's an acclaimed author of books and peer reviewed articles, often based on his research interests in the intersection of emerging technologies, such as artificial intelligence and robotics, with global politics and military innovation. Which is certainly something Michael, we're going to touch on today. Michael has also previously worked for the office of the undersecretary for defense, for policy in the department of defense, DOD. Michael, again, I'm delighted to have you with us. Thank you for joining.

Michael Horowitz: Thanks for having me.

Harry Kemsley: Good to see you. Now we have in recent podcasts, been doing a great deal talking about the power of open source intelligence and the open source information upon which it is based. We're going to start moving into the realms of how technology is being applied and how that's helping drive value and development of open source intelligence or OSINT, as I would call it. But before we go there, I'd just like to start as usual to level set the conversation about what the three of us mean by OSINT. So I'll start with yourself Michael, in terms of what you understand open source information and the intelligence drawn from it is. And Sean, if you could then amplify that, but then it comes from yourself. So Michael, what do you understand by the acronym, the term OSINT?

Michael Horowitz: It's a great question. I think about open source intelligence broadly, in that it refers to information frankly, that is not classified. So open source intelligence could include information gathered from commercial satellites, information gathered from social media, historical data collected on how militaries fight, that's then aggregated in a way that helps us understand warfare. It can include a lot of different kinds of things fall into the category of open source intelligence. I would just say, I think about it pretty broadly, is basically the things that governments are not keeping secrets about.

Harry Kemsley: Right.

Michael Horowitz: Which is a lot of the information available in the world. The newspaper contains open source intelligence, from a certain perspective.

Harry Kemsley: Yeah. Perfect. I totally agree with that. Sean, how would you amplify or add anything to that if necessary?

Sean Corbett: Yeah, I would agree the same. I think the important role there, is the sheer breadth of the information that's out there. A lot of people think OSINT as media specifically, scraping social media that sort of stuff, but it's far wider than that. I think the key things are that it's publicly or commercially available, you might have to pay for it, which is fine. I think the other key thing is the accessibility and the fact that anybody can legally get hold of that information.

Harry Kemsley: Yeah, publicly available information. And certainly Janes has been making a living out of this open source domain for 120 plus years. So I'm glad to see that the commercial part came out as well. Now in the description of this very broad, very diverse data set of open source information and the ability to draw insight into intelligence from these open sources, the very scale of it, the speed at which it's changing, the diversity of it has increasingly demanded technologies to become involved. And that's the primary center of this conversation today. So Michael, if I may, let's get started on this piece of how technology is starting to enable, really enable the value that's available to us in open source information and intelligence. And look at it in terms of how that helps the analysts. So what are those emerging technologies that you have spent so long thinking about and working on?

Michael Horowitz: So I think if you imagine open source intelligence as unclassified information, that helps us understand the world, then emerging technologies can play quite a role in helping us aggregate that information, since one of the big changes over the last several decades, is the sheer scope of information that is available. That's not just about social media, although that's certainly a part of it. It's also about all of the say, historical information that Janes has that can be digitized and aggregated or information that you get from commercial satellites. All that adds up to data points that you can use for collection purposes and aggregation. And that I think is especially where emerging technologies play a role. You could imagine artificial intelligence and the use of algorithms to make sense of enormous data sets. If you're trying to run models to understand sentiment on social media, or rather than human analysts, you're trying to look at every moment of commercial, satellite data, using image recognition algorithms to process that information and flag things that are potentially of interest for humans.

Harry Kemsley: Right.

Michael Horowitz: So I think that there are a lot of ways that emerging technologies and especially AI, can play a role in the open source revolution. And I think at the intersection of technologies here is especially important, in that it's not just about having an algorithm, it's about getting the data, storing the data, having an algorithm, interpreting the data, all these kinds of things come together. And note that I think I've said data four or five times here.

Harry Kemsley: Right.

Michael Horowitz: Data is really at the core of this.

Harry Kemsley: Right. Perfect. Sean, anything to add to that? I've got a couple of points, but before...

Sean Corbett: Yeah, Michael, you said that more eloquently than I could. But the way I see it's two main areas that it can help, is managing that huge data repository that you talked about. And I was looking at it through the lens of the analysts who still sadly has to do a lot by Excel spreadsheet. But if you sit and look at all the sources you mentioned earlier, you've got information coming in that's tailored to a specific problem set. It's just trying to just surely manage that data and bring it into one place. And then the second element of that, is actually using that data or being able to use that data in a way that answers a specific challenge or problem set. And the key to that of course, is when does the human get in the loop for that? But I agree totally, the amount of automation that you can do to manage that data, to get it to that place is absolutely critical.

Harry Kemsley: Perfect. Thanks gents. So I agree with both of you and I just want to grab onto a couple of pieces of that. I particularly like your earlier introduction about OSINT as being those things that are not classified, they can be available commercially or just publicly available. But I also like the link you've made there to the ability to do that aggregation and to find those key moments, where the analyst now needs to tune in, because there's something that maybe of interest found in the data source that may be vast and very difficult to otherwise penetrate. Because that leads us to that conversation I really want to spend some time on, is the so what, so we bring technologies in, it helps us do the aggregation, it helps us find the trends, the sentiment. But how's it helping the analyst because one of the things we're going to come on to later in this brief conversation is why is it not being more fully adopted in the department of defense? Why is it not being fully exploited in various governments of the world, actually not just the US. But to get there, I think we need to start looking at, well, what are the applications of open source intelligence to the analyst in their day to day role, whilst they do have all the exquisite technologies available to them, they have all the techniques and the analytics available to them? Why are they still not using it? We'll get to, but what are those uses that they should be using it for, Michael?

Michael Horowitz: Sure. I think that open source information can be useful for a variety of things to the analyst. And so if we're specifically talking about, all right, somebody's sitting in a government with access to classified information, what do they need open source information for? I think that there are a couple of uses, some of which are, I would say, inherent open source information and some of which are in effect of the way that many governments pigeonhole and silo information. The first thing is, and let me start with the second one, the way that governments silo information. I think one of the benefits of open source information is sometimes it can give you a broader historical view of what is happening, allowing analysts to take current events that they may have exquisite knowledge of due to classified information and place that in a better historical context, as they're writing reports for their bosses, for policy makers making decisions. Since, governments often excel at delivering information on say, what are Chinese missile capabilities today, or how have Chinese missile capabilities changed over the last year? But if you really want to understand how that should matter for decision making, it's good to have context on how they've actually evolved over a period of time and how the kinds of changes in capabilities we're seeing have mattered in thinking about conflicts in the past. And there's context there, that opens source information can provide. In theory, you could get that from classified information, but governments often don't aggregate it and store information like that.

Harry Kemsley: Sure.

Michael Horowitz: So it can be often easier to use open source information to get that, so that's one thing. The second thing, is classified information is limited to what governments are looking for. And often what happens in the world, is based not only on the crises that we know are going to happen or think might happen, but there are also things that we can't really predict and we think people might not be looking. So take COVID 19. There are a lot of intelligence reports from say, in the US, from the global trends and the National Intelligence Council saying," All right, there's a risk of a pandemic over the next 20 years." But when you get to the specifics of a particular pandemic, the classified sources weren't necessarily laser focused on Wuhan markets.

Harry Kemsley: Right.

Michael Horowitz: Instead, it might be open source information that especially initially before an intelligence apparatus is focused on there, brings insight and knowledge to the table. That's the second thing. The third place, I think open source can make a big difference for the analyst, is in the aggregation of unclassified information. And here, when we're talking about commercial satellites that might cover areas that intelligence assets aren't necessarily covering. Or say analysis of social media are examples of where you're taking data that is commercially available and processing that to try to understand how people are thinking and how people are feeling, which should be inherently helpful. And we're trying to understand as governments, how countries are likely to behave.

Harry Kemsley: Yeah. I love that. I've got a couple of points I want to come to you, but I'm going to hold myself back. Sean, go ahead.

Sean Corbett: Yeah, just very quickly. Again, that validation piece is huge, isn't it? The context and then you validate either with the classified sources or even vice versa, actually. And that leads to the second point, which is ubiquity. The intelligence community anywhere can't look at everything at all times and neither should it be, because it needs to be focused on specific priorities for that particular nation. And so inevitably, there'll be stuff that isn't looked at. Just the nature of open source now, is that wherever you're not looking at, if something happens and invariably things happen where you're not expecting them, you can immediately change and rely on that open source intelligence until you start skewing your intelligence collection in the right direction. So I think that's really important. And one thing is close to my heart as well, is it facilitates intelligence sharing. As long as you validate and trust the data, then you can use it and in unclassified format to share with allies and partners who you wouldn't necessarily want to share the classified stuff with, or some of the more technical collection means. So those are the big ones for me.

Harry Kemsley: Thanks, Sean. So I agree with again, all of that, I think for me, what's coming out of that, the precipitate from that conversation, is that open source information can save the analyst a huge amount of time. They can get to a good context, a good understanding of a topic by stepping into the open source arena, particularly, whereas you said, Michael, they don't have the legacy archive of everything they can reach to quickly. And indeed the drinking straw view of the world they often get, through very exquisite classified means doesn't necessarily give them that context either and that legacy. I think the other part that you've both touched on this ability to find verifiable information, also now starts to drive for me the second big so what, with open source in terms of application of the analysts and that is, that indicator and warning. You used a really good example, very contemporary example Michael, of if we had been looking at social media or medical records around Wuhan, maybe not withstanding the nation that it was in, we would've seen signs of problems starting to emerge, that the local hospitals are a little uncertain about and people are talking about between themselves. Is that now something the analyst should now poke and start to look at more carefully with their own more exquisite laser dot focus that they can bring with classified means? The third is the ability to join the dots. Again, it links to the aggregation, it links to the indicators of warnings and technology is particularly good at that. Certainly, Janes has learnt that lesson in recent times with our Jane's Intara capability, which is all about interconnecting the intelligence that we've gathered over a long time and the insights you draw from it. So those indicators and warnings, those insights you can get, the context for me start to summarize those big handfuls of what it is that the open source intelligence can mean as an application, that's valuable to the analyst. Now, conscious that time will evaporate on us if we spend too long on that, which I'd like to spend more time on, but I'm going to move us on. Let's now start digging into, so why isn't it being used more? What is holding the analysts back from using it? What are the policy issues we've got to crack? We've spoken about this with other colleagues from the US and elsewhere in the world, and there is a bit of a theme coming through. But I'm keen Michael, to under your experience in the policy area of the DOD. What do you see as being the reasons why the DOD is unable to scale OSINT within its capability portfolio?

Michael Horowitz: I think that you're broadly right. I would caveat by saying it's of course, the job of intelligence communities rather than DOD itself to scale open source intelligence. I think the question is what information does your average analyst in the department of defense use when they're assessing the world? I think there's a piece of this that's certainly the allure of classified information. Think about the reasons why information is classified. One, are things that are pre- decisional, you want to be able to have private conversations about government policy that you don't want to be public. And so those are classified for one reason or another. The second, is information that you believed is different than or better than what is available publicly. And that could be the difference between what somebody on the inside sees and what's in the New York Times. It could be different facts, it could be different confidence in the same facts.

Harry Kemsley: Yeah.

Michael Horowitz: But the notion of classified information as inherently superior, I think means that there's often a bias toward focusing on it. And again, that's not irrational.

Harry Kemsley: No.

Michael Horowitz: But the challenge is then when it leads people to only look at classified sources as a basis for making decisions, since of course, those resources are limited, what they examined might be limited, the angles that they're examining questions from, maybe limited. And so if there are ways to aggregate more information, it makes sense. So I think a piece of this is probably about cognitive bias, you have access to all these classified sources, shouldn't we be relying on them? And the answer is basically yes, but not exclusively.

Harry Kemsley: Right.

Michael Horowitz: And I think another piece is the bureaucratic politics of it in the, how do you create vehicles within governments to aggregate and validate open source intelligence and the use of it, in a way that helps analysts? In that people respond to the incentives that they're given. If a briefing that an analyst gives to their boss is more likely to be accepted, if it includes lots of references to classified documents, that if it includes a bunch of references to open source information, even if it says the same thing, then of course if you're the analyst, you're going to rationally respond to incentives. And so I think a piece of this is about the cognition in the individual and how we think about this. And a piece of this is about the bureaucratic politics of valuing and sending a clear signal about the valuing of open source information.

Harry Kemsley: And Sean, I know you want to come in on this, but just before you do, Michael, do you see there's a generational shift potentially going to happen? I look at myself as somebody who's had 35 years in this arena. 35 years ago, the idea of using open source information or intelligence for anything remotely close to decision would have been laughable. And my entire professional reputation would've been under question, I'm sure. More recently I've seen generations of younger analysts coming through, who are much more disposed to get in amongst the open source environment. Do you see that? Do you feel that as a possible shift in the way the cultural bias you talked about moving?

Michael Horowitz: I think possibly. I mean, one way to think about this and this analogy is not perfect, but just like we think about the locus of technological innovation as increasingly coming from the private sector, rather than from governments, something like advances in AI being potentially an example. Despite all of DODs funding of Silicon valley back in the day.

Harry Kemsley: Sure.

Michael Horowitz: But you could imagine something similar here in the open source intelligence world, where you have newer generations, more comfortable using open source information and more comfortable using different sources of information in combination, potentially to make decisions. I would also add though that part of it, there's also a shift in what needs to be classified. One example comes from maritime surveillance, governments if you wanted to track shipping around the world, you used to need to have people in every port, essentially counting, how many ships are there? How big are they? When are they leaving? Whose flag are they flying? Et cetera. Now for a small fee, you can access all of that information. And so I think a piece of this, is about information that used to be classified, migrating. You still need that information, if you're an analyst trying to work on the maritime environment.

Harry Kemsley: Right.

Michael Horowitz: But information that used to be classified now, can be accessed commercially. And so that I think then changes the reality, but also changes expectations.

Harry Kemsley: Yeah, good. Sean?

Sean Corbett: Yeah. There's a huge amount to unpack there and we just don't have time to do it. But the way I would focus on it and I agree with all of that, is part of it of course, is self- preservation. If you are a big intelligence organization and you need to justify your budget funding line every year, it's very easy to argue," Well, we don't need that if we can get everything from open source and how much you're spending on exquisite collection capabilities?" Now that's more perception than reality, but I'm sure it does come into the sort of thinking. And there are those definitely who still perceive that everything has to be super classified. I've seen an unclassified document turned into very highly classified, purely because the individual thought it wouldn't get read unless it was. Now that gets into more the culture, but that's not either surprising either, because if you talk... We'll hopefully talk about the next generation in a moment and I'm segue waying there as we are. But if you look at the tradecraft, if you are part of a particular agency, you've been developing your tradecraft for an awful long time, you know your stuff, you know what you trust and you know the efficacy of it and the value of it. If you coupled to that, the actual access, so believe it or not and it sounds counterintuitively, but it's easy to access stuff when you're sitting in a skiff somewhere at the high side, which is very highly classified, then it is to get unclassified stuff other than your CNN that's pumping at you all the time. So there's an infrastructure element as well. And then we've got to develop the tradecraft. I know that's being looked at now to think more broadly, how you do combine the unclassified with the classified? And that'll be my final point, is the aggregation of the data. We do have to be a little bit careful sometimes, because that can turn an unclassified piece of work into something that is classified, particularly if it gives away why we're looking at something and perhaps the intelligence gap. And that is again, a risk aversion. This is very natural, because it's drummed into the intelligence community from a very early stage.

Harry Kemsley: Yeah, let's just pick up on that point about tradecraft. Something that Janes talks a great deal about, is to deliver assured intelligence in the open source realm. We have a tradecraft developed over many, many decades, and we stand by that very, very firmly. However, do you see the responsibility for that tradecraft being a matter that the IC must retain? Or do you think they should be stepping outside Michael, into the commercial environment where there are I think, ample examples where tradecraft has moved on and has been enhanced, accelerated by the emergence of these technologies and the engagement with those technologies that's gone on in commercial environments? Do you see that as something that should become more partnership? Or should the IC retain full and total responsibility for that tradecraft development?

Michael Horowitz: To me that this is an empirical question, right? What helps us better understand the world? If what we're trying to understand, is what are the capabilities of the Chinese military today? Or how many launch tubes will China have in 2023? Or something like that. If you tell me that the... And exclusively I see focus will deliver a more accurate answer than an open source focus, then I might tell you something different than if you tell me that. And you're basically going to get to the same point either way. So to me, this is a little bit of an empirical question. Backing up to think about tradecraft, I'd say more if you want innovation, you have to be willing to look at what others are doing. And to the extent that organizations outside governments of which Janes is certainly one, have ways of doing business, gathering information, aggregating information, they can validate, then any responsible intelligence community should be always looking for ways to improve. And so, whether it's by going to those organizations, learning from them and implementing things internally, or partnering with them, you could imagine a number of different possibilities. But at the end of the day, the proof's going to be in the pudding, right? So to me, it's an empirical question of what works better. But I'd like to think that government agencies are despite being large and having lots of different stakeholders, should be looking for ways to innovate.

Harry Kemsley: Yeah.

Sean Corbett: Yeah. I would say that the IC tends to be quite precious, for good reasons actually, about its tradecraft, having developed it for a long time and following the policy that it has to. But I think you're absolutely right Michael, you've got to be able to amend that and adjust that. So I think they'll want to own it, but that doesn't mean to say they won't encompass and really engage with industry, because they are doing it to a certain extent, saying," Okay, how do we develop this?" Which leads on to the next thing, is what sort of analysts do we want? We had a big debate in DIA before I left and we had a great briefing from two people who had taught themselves coding effectively and they're able to use what they'd learnt and apply it to the data they had and come up with some additional, particularly trend analysis that we haven't seen before. And they were massively lauded for good reasons. I mean, they're very clever anyway, but then the debate goes,"Well, do we need somebody who is very good at the technical element? Or do we need somebody that really understands the problem set?" Now of course, the problem is the answer is both. And I think it goes back to your point, Harry, that is this a generational thing? Is it something that people of my age worry about, but I shouldn't be because the kids are all learning it? It's just that, yeah, whatever sort of thing. But that definitely needs some attention.

Harry Kemsley: Let's just touch very briefly, because I know time's getting short on that matter of data and data science and the prevalence of data science in many, many conversations we have today, around the intelligence arena. Now earlier on in this conversation Michael, you used the word data many, many times and we've enhanced that for sure in this conversation, since that. One of the things that I've identified in recent times, is that the tech Sergeant in the corner of the room, who's done the data science course, who really knows how to use Python for example. Just making that up, who can get his or her fingers deep into the data and start ripping out all kinds of trends and so on, starts to become the center of gravity for the analysis in the modern conversation. I've seen that plenty of times. Is that a healthy thing? As Sean said, do we need more and more" data science" into this arena, data wrangling I think is a more contemporary way of describing it? Or is there still the need for the generalist who can sit back and understand the bigger picture? What's your view on this data science aspect of this conversation?

Michael Horowitz: I mean, it's in some ways the easy way out, but I would say the answer is we need all of those things. And I would distinguish to start between being data literate and being a data scientist.

Harry Kemsley: Right.

Michael Horowitz: I think everyone, including generalists needs to be data literate. I would define data literacy as someone that maybe understands the difference between correlation and causation.

Harry Kemsley: Right.

Michael Horowitz: Someone that can easily work with tables and charts and information, who understands what statistical significance is, those kinds of things. And that helps you to be fair, understand the limits of data sometimes, as well as the areas when it can be useful. But I think that everybody needs that basic data literacy, because that in a data driven world just helps us understand the world around us. So increasingly, I think data literacies will be a key part of the toolkit that analysts and other kind of generalists that also subject matter expertise have. We also need data scientists, people whose comparative advantage is the ability to code or to take raw data and discern trends, however exactly you want to define, protect networks, however you want to define what a data scientist is going to do, is going to vary from organization to organization, even within organizations. And I mean that both in the private sector and thinking about military services or intelligence agencies. But we need more people that have those skills, because we do live in a data driven world, but that's not a substitute for subject matter expertise or for generalists that understand broad things about international politics, international security, military behavior, et cetera. It's a reason why everybody needs to be data literate, but we do need more data scientists.

Harry Kemsley: Yeah. Very good. Sean, what I'm going to do now, is I'm going to monologue for about a minute doing the summary of the conversation to date. But what I'm going to ask you both then, is if you could do just one thing and know that one thing would get done to really enhance the engagement, the exploitation of open source intelligence in the Department of Defense or the Ministry of Defense in the UK, what would that one thing be if you had the chance to wave the mythical magic wand? So while you think about that, let me just summarize what I think we've got to in this conversation. So we talked about open source information and intelligence as being the breadth and depth of so many things that are publicly available or unclassified. We've delved a little into the emerging technologies, such as artificial intelligence and how that's starting to enable us to do better aggregation, better identification of key facets within the data and even the interconnections within those data that are becoming useful for the insights they draw. We've also talked about how that can help the analysts do a range of things from seeing an emerging threat before it's become a threat. How that data application can save them huge amounts of time and with commercial support, some of that time can be really saved by using valued and assured content such as Jane's wanted to provide. But we've also looked at why perhaps OSINT hasn't yet been fully scaled, fully adopted by the agencies. And we acknowledge the reality, the rational reality about an analyst staring at classified information and just believing it more, because it is so much more readily verifiable, for example. And then we touched on briefly there about the tradecraft and the emergence of the data wrangler, the data scientists and how that's coming through as a necessary, but not complete aspects of tradecraft as it's developing. But nonetheless one that we have to acknowledge. So with all those things said then, I'm now giving you both a magic wand that only works once for both of you and you get to shake it once. I'm going to come to you first, Sean, on this occasion and Michael, you last as the guest, on what do you want to achieve, for OSINT to become the powerful enabler that we all think it could be, Sean?

Sean Corbett: Such an unfair question, but if I had to crosstalk.

Harry Kemsley: crosstalk.

Sean Corbett: I would say, it's the leadership embracing the whole concept of OSINT. There's a lot of good chat out there and a good," Yes, we need to do this." But the analyst won't change unless there is incentive to do so. So there has to be reward for it, there has to be a," I want this at an unclassified level or the lowest classification you possibly can." And then being seen to reward it. And by leadership, I mean leadership at all levels. If you can incentivize people, if it's," Right, what's in it for me?" They will do it. And that leaves aside the fact that it could be easier and it takes less time, all that sort of stuff, so that's the one for me.

Harry Kemsley: Well done, Sean. Michael?

Michael Horowitz: And Sean basically stole what I was going to say. So thanks, Sean.

Sean Corbett: Sorry.

Michael Horowitz: So let me add some texture there. I think one of the challenges, I'd say varieties of government actors, militaries intelligence agencies have with innovation adoption, broadly is when there's a gap between senior leadership rhetoric and the reality on the ground. And reality on the ground, meaning incentives as Sean suggested, budgets, et cetera. I think that if there's one key set of actors, that if you could wave the magic wand and persuade, that would be most important, it's actually less senior leadership. I think if you look at public statements, will tell you that they believe that open source information is important and it's not necessarily your analyst just on the ground who they're responding to incentives. And to your point, Harry, maybe there's some generational change that drive more openness. It's actually your mid- level management where their receptivity to open source information, them believing that it both will provide greater accuracy in reports and that they'll be rewarded for that accuracy and not rewarded for using open source information. I think that's actually, the key layer in varieties of government bureaucracies, that it's critical to get on board to accelerate adoption.

Harry Kemsley: Perfect. Now, given that you should never ask a question you're not prepared to answer yourself, I'll give you my answer and you have both taken the best answer. I would add to that, the point that you made earlier, Michael, about data literacy. If you think about it, if people are data literate, they will embrace the data available and understand the risks within the data, whether that's classified or unclassified. They'll understand the veracity of the data they've got in front of them, to the limits of the data they've got in front of them. And I think that will start to thaw out some of the permafrost layer that we are alluding to there, between the actual analyst doing the hard work and that is that's waiting to be supported. Now as ever with these podcasts, I know I could spend the next three hours discussing various parts of it, but we have run out of time, so I'm going to draw stumps at that point. Sorry for those that don't understand the cricket analogy, I've just used there, draw stumps, end of game.

Michael Horowitz: Yeah. I have no idea what you're talking about.

Harry Kemsley: Yeah, exactly. Right, okay. It's the bottom of the knife, it's the end. I don't know, you crosstalk.

Michael Horowitz: I'm an ugly American. I'm sorry.

Harry Kemsley: Don't worry. I don't actually like cricket. But moving on, the essence of this conversation as ever has been the ability for three people to come together and have a conversation about something that they all understand and that we can to draw insight from. For me, there are five or six parts of this conversation, I've got notes in front of me that I'd really like to dig into further. They'll probably become subjects for future podcasts. Sean, I'm just warning you of that. And Michael, if I may, I'd like to invite you back at some point, to start to unpack some of those things that we've discussed today at a relatively high level and drive down a little bit further into the depth of them. Because to be honest, it's only through these kinds of conversations that the audience, ourselves as well get to really understand the power of open source information and intelligence. And we hope will start to embrace it, whether that's from publicly available or commercial sources like Janes. Michael, thank you so much for your time today. Really grateful for that. And Sean, as ever, thank you for your contribution too.

Sean Corbett: Thanks guys.

Michael Horowitz: Thanks so much for having me.

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.


In this episode of the Janes podcast we look at the application of emerging technologies to enhance the value and to drive the development of open-source intelligence or OSINT. For example, we look at the use of artificial intelligence and using algorithms to make sense of enormous data sets. 

Michael Horowitz is the director of Perry House and the Richard Perry professor at the University of Pennsylvania. He's an acclaimed author of books and peer reviewed articles, often based on his research interests in the intersection of emerging technologies, such as artificial intelligence and robotics, with global politics and military innovation. 

Today's Host

Guest Thumbnail

Harry Kemsley

|President of Government & National Security, Janes

Today's Guests

Guest Thumbnail

Michael C. Horowitz

|Richard Perry Professor | Director, Perry World House