OSINT – What we learnt in 2022

Media Thumbnail
  • 0.5
  • 1
  • 1.25
  • 1.5
  • 1.75
  • 2
This is a podcast episode titled, OSINT – What we learnt in 2022. The summary for this episode is: <p>In this podcast Harry Kemsley and Sean Corbett revisit some of the key themes they covered in 2022 and discuss what they have learnt about the power of open source intelligence.</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: Hello and welcome to this edition of World of Intelligence from Janes. I'm Harry Kemsley and as usual my co- host, Sean Corbett. Hello Sean.

Sean Corbett: Hi Harry.

Harry Kemsley: So Sean, it's been a busy year. A year ago before the invasion of Ukraine by Russia, we were discussing a whole bunch of things that are with open source intelligence. And I thought it might be useful this week, this month to look at all the things we've discussed, not just in the last year but since we got this podcast started. Kind of a summary if you like, of our thoughts, our guests thoughts about open source intelligence. So what I thought we might do today, Sean, is let's look at why we think open source intelligence is important. Why should people be engaging with it at all or more if they've already started? Second, let's move on to the examples we've seen from our various guests about where open source intelligence has been applied, where it's been useful. And we've got a number of examples from that, as I recall from the various podcasts. And then finally, what are some of the considerations that we've learned from, again, many of our guests around open source intelligence that for somebody who's considering engaging or doing more with open source intelligence, they may wish to be thinking about those considerations before they get too much further. So if you're ready Sean, let's do a summary of the last year or so and let's start with the why. Before we get to the why, in fact, why don't we do just for the last time the definition. For those who listen to this podcast quite frequently, Sean, they've heard your dulcet tones describe in four parts, what is open source intelligence? Maybe we should print that on somebody's forehead so that nobody ever forgets. But what is in your mind the best definition of open source intelligence, Sean?

Sean Corbett: Yeah, thanks Harry. And I suspected you might ask me that as the first question, but initially we have covered an awful lot of ground over the last 18 months, two years. So I think getting it in 30 minutes is going to be fairly challenging, but I think we can do our best. But as you said, let's start with what we mean by open source intelligence. And the first thing, which I always say is a blinding glimpse of the obvious is that all the information used for open source intelligence must come from fully publicly or commercially available to all frankly. You may have to pay for it, but anybody on the street can access it. I think secondly and a slight nuance though is that it's got to be a obtained from legal and ethical means. That's certainly the law that we abide by. You can collect it but also apply it in ethical and legal means. So we don't use nefarious tactics or collection methods and we don't use it for bad things. I think thirdly, so the next two features are common to all forms of intelligence. Thirdly is it's got to be applied to a specific and specified problem set or a question and fourthly, and finally it's got to add value added. You can provide as much information as you like, but unless it answers that, so what, it's not really intelligence.

Harry Kemsley: Yeah, no, I agree. All right, well you mentioned the word ethics. I know we're going to come back to that later in terms of considerations. Let's just summarize your recollections, my recollection, Sean, of why OSINT is important. And just to get us started, I remember the conversation we had with Terry Bush where I thought he really, really well summarized one of the benefits of open source as providing context. If you have a highly classified source of information that gives you insight into a particular situation, open sources can provide a very broad context around that. He also mentioned indicators and warnings, that timeliness, that very high velocity of open source intelligence, if you can tap into it, can give you an indicator or warning of things. Again, I think we'll probably come back to that later in terms of applications. What are the uses, Sean, what other things about open source intelligence do you think make open source intelligence useful?

Sean Corbett: Yeah, I think the context word is really important and actually that does link to the indications of warnings as well because if you look at the intelligence specialist and how the intelligence community is actually organized, you get very specific areas of extreme experts who really know their stuff, they've been studying for a long, long time. But if you think about how many global challenges there are out there and you've got to pick up something that might be new or novel or even a different angle to getting that context, that is the why is this happening, what is the precursor to it, which trends are we seeing? Quite often you've got to go back into the more historical open source information, whether that's academia, whether it's just open source research or not, just to find out that exactly as you said, the context. And then if you look at that you can start seeing at a strategic level what has changed. Really big at the intelligence world. And if you're looking at something in a microcosm, you might just miss the fact that over a great period of time something has changed. Whether that's a political climate, whether that is capability development, I don't know. So I think those two are very important. I think the other thing that we do touch a bit on quite often is the ability that open source intelligence gives us to share intelligence. We always mention it, we don't go into any great detail, but if you want to hide your methods and sources, which of course you do, particularly when you are dealing with non- traditional partners who you know that within those organizations there are going to be a bad egg that you can't trust, the last thing you want to do is reveal your sources and your capabilities. So if you can disguise it, cover it by using open sources, open source sources, literally, then that is really useful tool.

Harry Kemsley: And I think alongside that, I guess you've kind of touched on it with indicators and warnings and context is that gap fill. You may well look at something through your incredibly capable asset that's giving you very, very high resolution perspective on a particular topic. But that context around it, that's the gap fill, that's the foundational intelligence that you may or may not have from other sources that open source intelligence can fill. And I think the last one that I can think of is that confirmatory piece as well. The ability to look at something through exquisite, perhaps classified means, get a picture that you think is right, and then to be able to confirm that, triangulate it with open sources I think is another use of open source that I've certainly seen when we've been working with clients out of Janes is the use of Janes for that confirm triangulation. All right, so one further thought. Go ahead.

Sean Corbett: Yeah, there is one other thing that I think is worth mentioning that again we allude to but not really mentioning specifically and that's reach, the reach of open source intelligence. Again, I go back to being very good in niche areas, but the world is a dangerous place and getting worse in many areas including non- traditional threats. But if you look, I was doing some research yesterday actually, that over 7 billion people on the planet have got mobile phones, 5. 2 billion of which are smartphones. If you look at what's on the internet right now, so there are 5 million terabytes, which I had to look up, on the internet right now of information, whether it's right or wrong, we can come onto that in a moment. And they reckon that by 2025 there will be 463 exabytes, which is times 10 times 10 more being published every day. I mean that is just staggering. Okay. And I'm sure we're going to come onto the okay, what do you do with that?'Cause it's too much. But if you look at that on a global basis, there are still spots that you can't get to. There are hard targets, the Irans of this world for example, and wild parts of Africa. But that just says that you can touch anywhere in the world with open source intelligence.

Harry Kemsley: And that also brings to mind that observation that Terry Bush gave us about if you are not engaging with this intelligence source, you're almost negligent in the world of analytics and intelligence.

Sean Corbett: Indeed, indeed.

Harry Kemsley: There is just too much value and I think there is a genuine challenge to come into later about how you access and exploit the value, but it's there to be taken. All right, before we move on then, let me just summarize what I think I've heard. So why would somebody or should somebody be interested in OSINT? Well, we've talked about context, we talked about gap filling and indicators of warnings, the confirmatory aspect and the fact that it's shareable and also the insights and maybe it's the insights piece that I'll just pick up on to drag us across to the next bit, which is where we start talking about where we've discussed in the recent year and before about the applications of open source and where it's been useful. We talked not that long ago, but I think it was sort of about this time last year about the coming of age of OSINT around Ukraine. And interestingly we are talking about an anniversary almost of the day of when the invasion occurred. We've been looking in Janes at Ukraine situation building up for two or three years prior. But nonetheless, whilst that was going on we were using open source intelligence tradecraft to track Russian movements, predict what they might be doing, et cetera. So the insight indicators and warnings that brought is one thing that I wanted to just take forward into the applications because when we spoke to Don Rasler back in March last year, Sean, you might remember we talked with him about counterterrorism and how open sources can be helpful. He talked about the fact that OSINT can mitigate strategic shocks, which is an indicator of a warning. The fact that you sense something is rising out of the open source sources and you can start to plot trend lines and begin to understand what that might mean strategically. There is a body of evidence to say that it was increasingly obvious that the Russians were going to invade and the war would start in Ukraine and I suspect many of the agencies knew that and were preparing themselves adequately. Certainly Ukraine seems to have prepared itself. But let's move on to the applications. Just mentioned the counterterrorism one, do any others stand out for you Sean, in terms of where we've seen open source intelligence really stand out as being a great source of intelligence?

Sean Corbett: Yeah, there's a couple, actually. One of them's very topical actually and I'll come on to it in a moment. But the other one, if you recall, not that long ago actually there was an attack, inaudible by the Russians with a missile that went into Poland. And very initially all the press got lit up. It's like, oh crikey this is the third World War, article inaudible the rest of it. But it was an open source analyst, very good one actually who I know, that looked at the imagery of the burning wreckage and when that is an S- 300 rocket engine and it was like, okay, who has S- 300? Well the Russians do funny enough and it's their defense, so do the Ukrainians. And very quickly from there was, ah okay, something might have gone wrong here. So that is immediate application of OSINT, not just the context and the strategic stuff, but then going back to the context as you know there's a raging debate going on about Leopard 2's versus Abrams and it's a lot of the information that's been used to why is there an argument, what is the difference has come back to, I don't normally plug it, but companies like Janes that say, well actually Leopard 2's are produced here here and they're exported. A lot of countries use them, they've got a lot of them, they're easy to maintain and they're also agile. Just what you need because they were designed for the third shock army and that's all come from unclassified sources. So you've actually got some decent commentary now, although initially it was quite interesting, why Leopard 2's as opposed to Abrams and that is all good context.

Harry Kemsley: Yeah, I think as well, just to touch on this point without setting off a rant for myself, it's interesting isn't it, how open source information can suddenly make everybody experts. I sit and grit my teeth sometimes when I look at how people are using open source data and describing it as intelligence when frankly as I suspect we've spoken of a couple of times, it's probably been taken from a newspaper that they've read on their way to work in the morning before they got to the TV studio. I think you've seen that in the recent weeks and months, have you not?

Sean Corbett: Yeah. And I think you know that's a niche of mine that I need to scratch because as you said, things are not always positive on the open source intelligence. And it goes back to the definition, add value and understand what you're talking about. You can have the data but if you don't know what it means and there are far too many, so this is a negative side now, year in instant experts on the television and without mentioning in names, there was one expert who was talking about Abraham's tanks the other day and you're thinking you've just lost all credibility there. But not only that, makes you wonder, well okay, so where did you get your data? Where do you get your data from? Anyone can read the Daily Telegraph and it's good, but you've got to have that multiple source and the context and the expertise to filter what is right and also what it means. So yeah, I think we're in danger of... It will close off though because eventually the instant experts will go away and those that really understand what they're talking about will continue.

Harry Kemsley: Well maybe they will, maybe they won't. We talked, did we not, and we'll come on to this in a moment in terms of considerations, the mis and disinformation that's out there. Abrahams may have been a slip of the tongue and that may just be misinformation as a result. It wasn't intended. But moving on, we talked about Ukraine, we talked about counterterrorism. I also remember you and I talking about the high North, the Arctic region and how open source insights about what Russia had been doing and more recently what we've seen from China and so on in terms of the exploration of that region now that waters are starting to become less frozen. That's another example, is it not of where open source over a period of time can, through appropriate tradecraft, start to build a really, really interesting picture and start to put together otherwise disparate bits of information. And if you remember, we looked at an airfield up there, I can't remember the name of it now, where a Russian deployment of aircraft in isolation appears to be of no particular consequence. However, when pieced together with multiple other bits of information we talked about at the time, there may well be a strategic intent there. In fact there almost certainly is. And if you remember that example that for me was a really good example of how context, gap filling, creates insights that otherwise might not have been seen.

Sean Corbett: Yeah, absolutely. And the high North as you know, is a particular interest of mine and it's a really good example because the whole point about the high North until probably a couple years ago was keep it demilitarized, don't talk too much about it, but we could see from an open source, whether it's commercial, satellite imagery, whether it's communications, that there is and has been a buildup by all sides in terms of military capabilities against all the reasons that we know, new trade route, the hydrocarbons, the rare earth metals, all the rest of it. And belatedly it has started to peak the interests obviously of certainly the Arctic countries and of course not just Russia but China, even India, people like that. So it is now becoming of great interest, but understanding the environment which is still quite demanding and then how to counter what might come out of there. And you've seen, and in fact Janes has done quite a good report, good load of reporting on it, the US Coast Guard have now got a new strategy for the Arctic, the UK are developing capabilities that are more useful for the Arctic. So we are seeing a potential militarization there, but taking that in context and of course a lot of nations, particularly the Arctic nations themselves, this is our backyard. So kind of leave it alone, but I think a lot of that can be covered on the open source domain.

Harry Kemsley: Yeah, I think you've underscored for me there is the open source environment, it's longevity, it's endurance in terms of the way it collects information and its availability means that if and when one of those backyard countries really steps into that for analysis, they don't have to have relied upon multiple exquisite capabilities to be looking at the Arctic. They can actually go to open source to get themselves that foundational understanding and indeed a bit of context before they do, although I'm quite sure they are looking at it very closely, as you said. Turning it from the more national security related. Do you remember we spoke with Claire Chu from Janes with Intel track capability, which is the geo- economic statecraft, the idea that transactions in the commercial space may not be just about commercial endeavor, it may actually have a national security purpose. I thought that was fascinating because what it did for me was underscore that you can get to actually quite sophisticated, in fact quite niche levels, of insight into activities and interpretations of activities by piecing together open source intelligence. Do you remember that podcast?

Sean Corbett: Yeah, absolutely. It was a really good one and it's something that I don't think even now is being focused on enough. I mean at the end of the day and you're seeing it again in Ukraine and to extent in, no great extent actually, in China that it's about the economy. So the ability to sustain yourself in any conflict, whether that's a frozen conflict, whether it is a political conflict is absolutely key to success. Without that you can't. So the economic state craft is so important and we really should be focusing on that far more in my view, particularly in the area we've just talked about, the indicators and warnings to discover the intent as much as the capability. And they talk about war being, totally misquote Klaus Fitz, war being the continuation of policy by other means. Well policy is about economic tradecraft, that's the basic tenant of a state. So understanding what nations are doing, and in this case I know it was focused on China, quite rightly, why they're doing it. You can really get into detail by looking about what they're doing with their economy and how they're trying to influence other economies.

Harry Kemsley: And that on insight, which I say is based on fact insofar as that if you are now trying to work with a Chinese or Russian organization, commercial organization, and you now are told that that same organization in a different part of the world's been associated with difficulties be that, some sort of slavery issue, some sort of local contamination of the environment issue or something more serious in terms of security. Well that would be useful to know. And that's the kind of stuff that's available readily from open sources. But if you didn't know to go and look for it, you might never know until it was too late and they are quote, " in your backyard." I also remember quite recently, Sean, this probably links us back a little bit to the insights and the coming of age of OSINT, the conversation we had with Warren, Warren the journalist from the Wall Street Journal, which for me, him talking about his relationships with certain US agencies, which he has to report on as correspondent for the paper, telling him that... Telling us sorry, that he had achieved a sort of 90% accurate picture from open sources in his responsible journalism, I thought was quite a telling statement, quite a telling statistic that he's saying his assessment and the assessment of the agencies he's working with is he's getting very close to quote, " ground truth." The kind of ground truth that we thought was only the purview of the classified agencies.

Sean Corbett: Yeah, that was a good one actually. And for me, the key to that was objectivity and the back trust. And that goes two- way if you've got a responsible journalist, and I think I was quite rude about unresponsible journalists at the time, but he was very much so investigative, wanting to get to the truth for its own sake as opposed to spinning it, sensationalizing it, or even putting a political bias on it, that's got real value. And I think the people are not going to leak information they shouldn't, but getting that trust in what they're trying to achieve and how they're trying to achieve it is really important. Getting to the ground truth as you say.

Harry Kemsley: Yeah. I do wonder after God willing, this conflict in Ukraine ceases that when the war crimes are being investigated, perhaps on both sides, that the open source environment will probably be one of the better sources of evidence to support the investigations that will necessarily have to go on for justice to be found. Just picking up with that word responsible Sean, leads me towards some of the considerations that we've discussed with guests in recent time. So before we get to that then, so what we've said is open source intelligence is important in so far as a number of things that it can do in terms of its value, context, indicates of warnings, confirmatory, it's shareable. We've talked about how the insight from open sourcing journalism has been effective in places like counter- terrorism, in geo- economic statecraft, the high North, we missed one actually the Special Forces community that we talked with Glen.

Sean Corbett: Yeah, Gwyn.

Harry Kemsley: Gwyn. An interesting perspective he brought about mission risk tolerance and the need to get things done in no fail missions and if necessary, open source was the basis of the intelligence that they would have to work upon. I thought that was an interesting perspective and indicates that they have a higher risk tolerance than many organizations might because of the nature of the role they've got. But the fact that they still saw the value and would act upon it for me was another underscore of the applications of OSINT.

Sean Corbett: Yeah.

Harry Kemsley: Moving on then from applications then into some of the considerations. So let's go to the first one being the technology because probably it's technology. I think we had with Emily, wasn't it back in May of last year, Emily Harding from CSIS, she talked about the coming of age of technology that is enabling OSINT, that was the Oscar.

Sean Corbett: Yeah, that's right. Yeah.

Harry Kemsley: Welcoming Oscar. She talked about the three revolutions in data you mentioned earlier, I can't even remember the name of the scale. Did you say Zetabytes or Zetabytes?

Sean Corbett: Oh, crikey.

Harry Kemsley: Big number. I can't even consider the size of that number.

Sean Corbett: Huge amounts anyway, yeah.

Harry Kemsley: Lots and lots of zeros. The vast quantities being one of the revolutions, but the availability and security of the cloud. So all this information is out there to be plundered, but also now the applications of AI and machine learning, those are the three revolutions you talked about, if you remember Sean.

Sean Corbett: Yes.

Harry Kemsley: And I think those are the things that are bringing OSINT into everyday speak and everyday part of national security.

Sean Corbett: Exabytes. Well, I've just remembered that. And that's a key and it's a thread that we're going to talk about I know again very, very soon actually. But A, you have to consider all of the information as we've already said. But how on earth do you manage that? You certainly don't do it on the old Excel spreadsheet. So you have to embrace, develop the technology that allows us to do that. And then that starts to take into the conversation, okay, so it's the data versus the sort of softer, the actual analysis, the trade craft. How do you cohere those to get the right answer when particularly whether you're in time constraints? So it's filtering the information in a way that still allows, again, and the phrase I use a lot, the human in the loop, but at what stage do you get the human in the loop and how do you evaluate the algorithms that are using to sort the data in the first place? That is the key question for me. So data versus tradecraft is such a key element of the conversation.

Harry Kemsley: Well, I think we agreed when we spoke with Emily though, that these three revolutions in the data are demanding change. To go back to Terry Bush's point about you being negligent for not to try and exploit the potential of open sources and there's that increasing, I think Emily described it as the urgency for change, that there really, really is a pressing need for the agencies to really grapple with it. But that, as I recall, she was just saying, there just isn't enough pace in the adaptation of it. I think we should probably spend a couple of minutes then Sean talking about, can you remember what it was that she said about why that be? Why do we think the agencies are not actually engaging with it as quickly as we could? What are the impediments? And I think the first one that I remember anyway was the culture, the concern that open source is full of myths and disinformation. We'll come back to that in a moment, I suspect and risk. But there were many other things that she talked about, if you can remember any of them.

Sean Corbett: Yeah, there were. But I think the cultural one was an absolutely critical one to be honest. That getting the, again, you've got to put your yourself into the sort of mindset of the analysts. They're so busy doing their job, getting the reporting out, making sure their standard trade crafters, right? Getting the terminology right, answering the question that, oh, here's something else to consider. By the way, make sure that you're using all this extra information in a technological manner. Now they can't do it independently and equally, and we've all seen the whole procurement process both within the UK and US and everywhere else, I guess, which is clunky. It's just not agile yet. So when you are talking about technology development, which is happening at incredible pace and learning as you're doing, how would you inculcate that and incorporate it into your trade craft? And that's almost impossible. But one thing, other thing she alluded to was the policy. The policy does not catch up with what it is you're actually trying to do. So you're still being risk averse because it's all about the security. You're still working out how to work effectively in the cloud, and that was another one of the things that she was talking about, in the secret domain or highly classified domain whilst also trying to incorporate unclassified stuff. Now how do you A, smash those together and B, in a way that the technology can use them both at the same time or at least cohere them together. That is incredibly challenging for anybody. So it is a really difficult problem set. So it's not just the culture, it's not just the policy. It's actually quite difficult to do.

Harry Kemsley: Yeah, I agree. I suspect we should come back to the technology. In fact, as we come on next to mis and disinformation, one of the things that I recall being discussed was we often use the acronyms, AI and ML, machine learning. We don't often talk about machine unlearning. In other words, finding out that we've been mis or disinformed and we've allowed that to feed itself into the machine at machine speed and yet we don't have as much dialogue it seems to be about machine unlearning, unraveling the mis or disinformation.

Sean Corbett: No, that's right. And again, we did cover that, didn't we? And some of the disinformation, misinformation out there is quite scary actually. In fact, I don't know if you saw recently, there's just a bit in the news right now where they are, they've called it the new AI spitting image. So basically they're getting famous people who are doing sketches and skits. It's not actually them. It's just been recreated through artificial intelligence. You cannot tell the difference between them. You really can't. Now, when we start getting into deep fakes, at what stage do you A, counter the misinformation or disinformation, well, it's both. And so exactly as you've said, is there as much effort going into the algorithms that enter the disinformation or unlearn if you like. I don't know what the answer to that is, but I know it's quite challenging.

Harry Kemsley: Who was the lady that we had on about mis and disinformation?

Sean Corbett: Yeah, I'm just trying... She was absolutely excellent.

Harry Kemsley: I remember while I was struggling to remember that, what I do remember-

Sean Corbett: Di Cooke. It was Di Cooke. She was great.

Harry Kemsley: Di Cooke. That's right. Di Cooke. Sorry, Di. My bad. What I do remember the nub of that conversation being is that the rate of improvement of AI generated deep fakes is such that we're struggling to keep up with the counter AI deep fakes technology and that the war of AI that's going on, isn't going well so far. I also remember talking about the fact that we assume, don't we, that us humans can tell the difference and we can look at something again that's not real or indeed hear a voice and say, " That's not really that person," et cetera. And she's saying that from her research, that's actually not true. We're not very good at all us humans at spotting mis or disinformation, whether it's deep fake imagery or indeed some text, which is quite a worry, isn't it, when you think about it.

Sean Corbett: Yeah, exactly. Right. And actually it's getting worse because the deep fakes are getting better and we are still the same old human beings. And I think because she was doing some really heavy research on it, I think that must be due to be published soon and I don't think it's going to make great reading.

Harry Kemsley: No, maybe we could get dive back on soon just to find out where that's got to. All right, now we talked about the use of technology and the necessity of that use. It's not a matter of debate, it's not a desirable, it's an essential part of dealing with the vast amount of information and the speed at which it's changing and being updated. That for me brings in Dr. Amy Zegart's podcast that we had about the ethics of the open source. You remember that was quite an interesting conversation in itself in that of course, we who have been and around the intelligence world for a while kind of assume that it's perfectly fine if somebody's left their privacy boxes unticked, that we can plunder their Facebook page or such. Not so much. And you talked in your definition of OSINT about not doing nefarious things well, defining nefarious and keeping ourselves ethical was the topic of that conversation, wasn't it?

Sean Corbett: Yeah, it was. And I remember you saying that you could be bribed with a pizza as well, but for that... So yeah, we get into some in interesting grounds here. I'd like to think we take an ethical approach anyway, but as Amy said, who decides what's ethical? It's like doing what is right. Well, who says what... Which doesn't necessarily mean legal. I was re listening to it the other day actually, but who decides what is right and what is good? I don't know. So I'd like to think we take an ethical approach anyway, but it's as much as about the application as it is the collection. So it's the full intelligence cycle if you want to call it that. But my view is if you're stupid enough to publish it in social media, then it's fair game as far as I'm concerned. If you manufacture that sharing of the intelligence in a different way, then it is not. That's a very black and white perspective from me, but I think it's good guidance. But the other thing she said, which is really important is you can't just do an analysis, write a report and then go, right. What's the ethics consideration? It has to be in Amy's words, baked in.

Harry Kemsley: Yeah, I think that's right. And the technology piece that she covered, AI in particular, and the sort of black box, the explainability, potentially separates the analyst from understanding the imperative and the consequences. The imperative for the analysis is whatever the mission is that they're supporting, but the consequences of their research, their analytics, may not be connected quite the same way when they don't actually know how the black box has done what it's done. I mean the human in the loop or human on the loop discussions, we probably don't have time for today, but I don't think we need to overlook the fact that intelligence is more than just analytics. It's also a process of persuasion. We're actually doing decision support and we need to be doing that for good, but defining what good is, is challenging. Do you remember what her thoughts were about what we should be doing about ethics? Because I think we got to the point where we agreed, yes, okay, it needs to be a part of the calculus. Can you recall, Sean, about what we might do to mitigate the ethical issues within the open source environment?

Sean Corbett: Yeah, so one of the things she talked about is just raising its exposure, which I think is what we've done. But the other thing is it needs to get into training. It needs to be trained as part of the overall process because she did call it a process more than anything else, which is going to be quite challenging in itself. So she had sort of three or four, didn't she in there.

Harry Kemsley: I think you're talking about the inaudible needs to create a explicit guide on where the boundaries might lie. Now that might be policy driven, that might be even legally driven, I suppose for some aspects of the ethical use of intelligence, open source or otherwise. But I think for me, it hinged on if the analyst isn't even thinking about the consequence of the analytics and the sources that they're drawing, then there is great danger for ethical inaudible sources. I think that's probably a-

Sean Corbett: And taking that to its natural... Sorry, taking that to natural conclusion, I think what she also said, which I think you guys are doing actually, is have an actual, for every organization, it needs to have an ethical policy. So you need to consider it as part of what you do rather than, as I said at the end, oh, we'll just consider it. And that starts at the beginning, institutional culture, if you want to call it that, everything from training to be indoctrinated and everything else, which is again, a challenge. But every single OSINT provider should have an ethical element to it.

Harry Kemsley: Certainly for commercial providers who can make choices about who they sell to and what they sell for what purpose. And certainly we do at Janes have a ethical policy that we try to apply to what we do for whom and for what purpose. The next one I want to just bring us into, because it kind of leads, I suppose on, it was one of the subsequent podcasts we did was with Dr. Claire York. Empathy. Now both of us had to go and find the dictionary to work out what empathy was. But I think we both walked away from that conversation, Sean, thinking, yes, I can see how an empathic view of what I'm doing or what I'm seeing might help me take a slightly more balanced view, a more balanced perspective.

Sean Corbett: And I know we're going to come back to this. In fact, we're going to bring them all together in what should be a really fascinating piece, which talks about the technology, the empathy and the ethics. But for me, the empathy, and I look at it again from a very pragmatic perspective, I want to get the best analysis out there, and if that means that I need to remove unconscious or conscious bias by putting myself into the adversaries' mindset, their view, thinking. So is that empathy? It is because it's understanding the culture and how they're thinking and why they're thinking a certain thing. But that can only be a good thing because it gives you back to the, okay, they may have these capabilities. What's the intent and what is driving that intent, which is such an important part of the intelligence process. And it's something, again, you mentioned it earlier, that could be very, very valuable for the decision makers. So why are they doing this as opposed to just what are they doing?

Harry Kemsley: And trying to look at it from their perspective. The motivations we might look at through our own prism that is terrorism, one man's terrorist is another man's freedom fighter as they say. I'm not suggesting that we forgive what we believe to be terrorism because it's somebody rebelling against something. That's not what I'm saying. But understanding it through the perspective of the actor might give us a different perspective and therefore a different reaction to it. I seem to recall Claire quoting the post 9/ 11 moment and the empathy that might have been required there. I have to say I have difficulty with that, but I do understand the point she's making. I am empathic to that perspective that we need to have a view about not just our perspective of what's going on, but the motivations and why those motivations might be as they were given the actor and their circumstances. So Sean, what I think I've heard then in the last... Goodness me, inaudible 40 odd minutes or so, is we've talked about the power of open source intelligence and it's utility in a number of applications. And now we've added to that the considerations we have to have technology, we have to understand the realities and dangers of mis and disinformation as well as the ethical considerations. And this additional one we just added, which is the need for empathy in the overall approach. Starting to make me feel like a really difficult job, analytics in national security is getting more and more difficult. Would you agree with that?

Sean Corbett: Totally. And I thought you were going to say, what's your last takeaway and so-

Harry Kemsley: I'm coming to that.

Sean Corbett: Okay. Oh, well. Okay. I'm not going to say anything. No, I will. I think what it underscores for me is that open source intelligence is every bit as difficult, and in fact, maybe more so than traditional intelligence, it is not an amateur sport. This is for professional people who understand what they're doing, why they're doing, how they're going to do it, and apply some real expertise to do so. We're all armchair experts, aren't we? Having read an article on something, but it's a lot more than that.

Harry Kemsley: Yeah, I agree. We talked a long time ago about one of the takeaways, one of the last thoughts was the need for data literacy. And I think for me, and I am going to come to your one last thought before we finish Sean in just a second. For me, one of the things I really want to underscore is that need for data literacy that allows you to look at data somewhat objectively and understand it for its potential fallibility, mis and disinformation. Does that really make sense? And be a bit more skeptical about what it is we're seeing? I find the armchair warriors, which you and I both are these days, and the belief in what I read in social media or other media forms or other open sources, that bit has to become better. We have to have a more generally higher level of data literacy. And your final thoughts take away then. What's the one thing that remains enduring as key to any intelligence, but certainly in the open source space from your mind, Sean?

Sean Corbett: So the more I think about it, as I said, the more complex this becomes, but what we can't do is become overwhelmed by it. The huge expanse of what we've talked about and we say we're superficial, but actually we're starting to get into some detail here. That a lot of that, particularly bringing it all together, the complexity is really hard. But at the end of the day, what are we trying to achieve with that open source intelligence? We're trying to get as close to the ground truth as we possibly can back to the definitions and add value. And as I keep saying, intelligence is your best possible assessment of what you think is going on in the time that you need to do it to best inform the decision maker. So if you wait for a hundred percent of the intelligence, it's never going to happen. And if you try and cohere all the elements in such a way that it's irrefutable, you're never going to get there either. So it's trying to take a pragmatic approach while considering all this myriad of issues.

Harry Kemsley: For me, I think that's all built upon an understanding that tradecraft, which is the fusion of great process, great judgment, great sources, those three things brought together even in the modern day where the sources are exploding in their volume and velocity, et cetera, there is still a need for the human in the loop to do the things that machines can't do and the machines can enable. But if you don't update, adjust and think about your tradecraft on a constant evolutionary basis, you are going to get overwhelmed. So for me, tradecraft, tradecraft, tradecraft, I can't remember how many times, but it'll be tens probably hundreds of times did we talk about that eventually it's going to fall back to Tradecraft. Eventually that's what it's going to be. And I think that's something we need to constantly update and evolve in this coming of age of the open source intelligence and its power. Great. Well, that was a very quick run through of hours and hours of podcasts. I think, Sean, what we need to do next going forward now is maybe continue to look at the applications of OSINT, look at the considerations. But now I think that there are quite a few things in there we need to revisit. Maybe next time let's bring Emily and Dr. Claire and Dr. Amy together with one or two others. Let's have a proper look again at that technology ethics and empathy thing. 'Cause I think that's going to be really challenging. Let's do that soon.

Sean Corbett: That'll be incredibly challenging, but also a lot of fun.

Harry Kemsley: Yeah. All right, Sean, thanks inaudible. Good to speak to you. Thank you.

Sean Corbett: My pleasure. Take care, Harry.

Harry Kemsley: Bye-bye.

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 podcast Harry Kemsley and Sean Corbett revisit some of the key themes they covered in 2022 and discuss what they have learnt about the power of open source intelligence.

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