Review of 2023

Media Thumbnail
  • 0.5
  • 1
  • 1.25
  • 1.5
  • 1.75
  • 2
This is a podcast episode titled, Review of 2023. The summary for this episode is: <p>Over the course of 2023 Harry Kemsley and Sean Corbett hosted more than 20 podcasts exploring topics such as the implications of AI for OSINT, the need for empathy in</p><p>decision making, and tradecraft in open-source intelligence. Janes analysts also discussed how OSINT has supported Janes analysis of new and emerging situations such as that in Sudan, Haiti, and Israel.&nbsp;In this episode Harry and Sean look back on the key themes and what they learnt from the discussions with their guests in 2023.</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 Jane's World of Intelligence. Unusually just the two of us, Sean. Hello, Sean.

Sean Corbett: Hi, Harry. Good to be back.

Harry Kemsley: Good to be back. So, Sean, 2023, 20 episodes, multiple guests, today just the two of us as I've said. As we did last year, I think it would be quite good for the two of us to go through all the things we discussed and draw out, I don't know, four or five of the key takeaways that we've learned from all the various expert guests and discussions we've had. I'm absolutely certain we could probably spend the next three or four hours covering it all, but let's try and keep it relatively brief. Of the things that we've discussed, let's try and find, as I said at the moment ago, four, five topics we really want the audience to highlight. And as we do so, we'll be saying" Thank you" in an indirect way to some of the guests that have been on as well. So where do you want to start? What are the big takeaways for you, Sean?

Sean Corbett: So I think first up to say that having done the research on terms of what we covered, we covered a huge amount of ground, and it's going to be-

Harry Kemsley: Sure.

Sean Corbett: ...quite challenging to come up with four or five, and we've been really lucky to have some great guests. So, within that context, I'm sure there will be people we miss out, but I think just literally the richness that we've got, well, that's going to make that necessary unless inaudible-

Harry Kemsley: Absolutely-

Sean Corbett: ...yeah.

Harry Kemsley: ...and how many times this year and previously have we said at the end of every podcast, " We really need to come back to that"?

Sean Corbett: Indeed.

Harry Kemsley: And maybe at the end of this one, we'll talk about the ones we are going to come back to later this year. Back to you, Sean.

Sean Corbett: Yeah, sure. So, if you recall, I think it was two years ago that we actually came up with the term, " the coming of age of open source intelligence." And I think at that stage we were looking at the fundamentals, foundational level in terms of how it was being applied, and how rich inaudible was becoming. But I've got to say that since that time and last year in particular, I think some of the episodes we covered just demonstrate that we've gone to the next level in terms of applications and in terms of the coming of age. So I think my first point would be that while that foundational level intelligence is so important and critical, I think the ability now for the responsible open source intelligence community to actually discuss and provide value on current intelligence and even emerging crisis that happened at speed and with accuracy, I think it's really, really impressive actually. And if you look at thinking back to what we did in Sudan, we were slightly ahead of the game in terms of the reporting on that. And in a very unsure environment, which there was lots and lots of conflicting information, and not being a particularly big collection opportunity, we were able to get into granularity particularly quickly on that one and that was Sudan. So that was one that really brought inaudible.

Harry Kemsley: Yeah, I think that's right. If you think about, as you were just saying, Sean, in your introduction there, we've often talked about open source intelligence as being that foundational layer, the gap fill to our understanding of a situation or a particular piece of equipment or whatever it might be, what you're saying is, using the Sudan as an example, actually the open source environment is increasingly relevant and useful to the current intelligence understanding, understanding what's happening right now, getting those insights. And doing it at speed. I think that alongside what we did with Lewis and Jake Shustey around Iran as well, there is an underlying sort of indicators and warnings piece that came out of that podcast back in March where we had the conversation about what we can see through open sources in a very closed environment, actually, with very limited collect, as you were just saying about Sudan, limited collect, but still able to derive insights that have current and potentially predictive qualities as well.

Sean Corbett: Yeah, absolutely. And I'll go even further actually and say that the piece on Gaza, if you recall, where everybody woke up to, " What on Earth happened there?" And-

Harry Kemsley: Yeah.

Sean Corbett: ...two of our very proactive and extreme professional analysts were way ahead of the game. They were reporting stuff that had not yet been validated and verified. That I from what I understand was extremely useful in triaging for the intelligence community to say, " Right, what have we got so far? Oh, here's some-

Harry Kemsley: Yeah.

Sean Corbett: ...stuff done." Now there's a really important point, and I guess this thread goes through all, you can only do that if you are looking at something in advance, and that you've got that baseline understanding to say, " What has changed? What is different? What are we looking at here?" You can't just come from scratch with a blank sheet paper and say, "Right, what's happening?"

Harry Kemsley: Right. So it's the combination of that foundational long look that allows you to identify what's relevant and particularly important in the current look that's going on. I do remember with that conversation we had with Emale and Lewis around the emerging situation in Gaza as it was at the time. We had had the fortune, had we not, that Emale had just come out of that locale, and had some direct insights from very, very recent times. There's a bit of luck there. So that's not to deny that, but at the same time, because he had been studying that part of the world so long, he was able to identify the key aspects for the current situation that as you say, allows us to triage what's going on and where we need to be looking. But I think there's also then this question mark isn't there about if we can look at open source intelligence as being that foundational layer and now we're increasingly seeing the coming of age of the capability, the more current, this predictive piece is probably a bit we should come back to in the future, but looking forward into what open sources might be telling us might be coming around the corner. That absolutely requires us to combine current historical foundational intelligence with tradecraft to look forward. But I think that's one of those areas we might want to look at in the future.

Sean Corbett: Undoubtedly. And just to underscore that actually, talking to some of my community friends in the US in particular, it is going to be the next big thing. I mean you know I've always said, the Nirvana is predictive intelligence-

Harry Kemsley: Sure.

Sean Corbett: ...which is always the most tricky obviously, and there's so many things that could happen, but I think getting into that world of, " Okay, these are the feasible things that might happen, this is what we're predicting," not all of them will come true. But at least if you're looking ahead, you've got a chance of either reacting when something does happen or even stopping something from happening by saying-

Harry Kemsley: Absolutely, yeah.

Sean Corbett: ...inaudible at this part of the world because something could happen.

Harry Kemsley: Yeah. And I remember the conversation we had with Matt and Lewis from Janes about Haiti, and we talked about how the combination of commercial and potentially government agencies working together to look at what is on a global scale a relatively low priority for some parts of the world for sure, but the ability for the commercial world to work in partnership with the government agencies allows us to look at that. And in that particular case, again, it was that predictive intelligence element that came through as I recall, that indicators of warning space that we talked about then back in the autumn of 2023. Well-

Sean Corbett: Yeah.

Harry Kemsley: ...look, one of the things that-

Sean Corbett: inaudible just while you talk about Haiti-

Harry Kemsley: Yeah.

Sean Corbett: ...slightly aside, but that was a really interesting one for me because it talked about substate activity.

Harry Kemsley: Yes.

Sean Corbett: And the granularity in the detail you need to go into that is quite extraordinary. I'm not saying that you don't for other things, but to understand what is an incredibly complex, but-

Harry Kemsley: That's right, I remember that now, yeah.

Sean Corbett: ... anon- traditional intelligence challenge, and yet we were still able to address that in a fairly comprehensive way. So maybe that's another element that we should look at in the future in terms of non- state, we've talked about non- traditional threats, but whether that is people smuggling, whether that is organized crime, et cetera, there was some real potential there, I think.

Harry Kemsley: I don't remember the number, but I do remember the point that Matt and Lewis were making about the incredible number of diverse substate, non- state actors that were operating in that country. And as you say, an ability in the open source environment to plunder the various open sources to get an understanding of those multiple groups and actors was a key piece probably of the foundational intelligence an organization could get if they were going to go into Haiti, for example, as a non- state, non- government organization for example. There would be all kinds of support provided there. Throughout all of this, one of the things that certainly came out though was the need for us to engage with technology. We heard on a number of occasions how technology was becoming not a nice- to- have, it's an absolutely essential part of dealing with the information environment that we live in. And I do remember the conversation we had with Keith, Keith Dear around the use of AI in support of the open source as well as of course, other intelligence work. I think for me, one of the things that came through there, it wasn't just the fact that these advanced technologies are able to help us in terms of the collect and to some extent the collate, the ability to deal with that at scale and at great speed of course, but also that as I recall, we looked at it as being a necessary tool but not a sufficient tool. It wasn't one of those things that we decided AI was going to answer all of our questions and that we needed just to sit back and let the black box do its thing, wait for the answers, and it would provide us the answers. It was very clear to me from that conversation as well as others we have with the people like Randy Nixon from CIA, that this blend of technology with tradecraft, good judgment, et cetera, was the key. If you remember with Keith in particular, he talked about it almost being negligent if you didn't use these technologies to do the collect. And that that mundane automated collect piece was really where AI was really still very, well, not still, is very, very much in charge, if you like, because it can do it at scale and at speed. But the human very much still needed to be there at the other end of that intelligence cycle, which was and the more theoretical and conceptual part of the process, the judgment and analysis.

Sean Corbett: Yeah, I thought the Keith Dear one was excellent from, it was very balanced. I think it recognized that everybody talks about AI, if you don't put AI in a conversation now you inaudible clearly inaudible-

Harry Kemsley: Not the whole thing, yeah.

Sean Corbett: ...inaudible anymore. And then you get the sensational side, like, " We're all going to get killed by robots and the world's going to get taken over," et cetera. But I thought that gave a far more balanced view on it's a tool, use it as a tool. And yes, of course you've got to have ethical AI and we've got to make sure that it's still the same control. But as you said, everything from just sorting the huge amounts of data out there as a tool for the analyst to make them better, up to the more right- handed of the spectrum where they are helping you join the dots together. But I still, and many people have said this, I think as well, I still say that you cannot right now anyway, and maybe not for some significant amount of time, replicate the cognitive processes of human being. And of course that is partly the analysis and partly the tradecraft, and Neil talked about human machine teaming and it was absolutely inevitable you had to use it.

Harry Kemsley: inaudible.

Sean Corbett: But-

Harry Kemsley: More widely, that's right. Yeah.

Sean Corbett: ...within the context of, this is how to help an analyst do their job. And then segueing again, there's the ethical element which how ethical can AI be? It is garbage in, garbage out. So at the end of the day, yes, of course there is developmental AI and unexplainable AI, et cetera, et cetera, but there's still an ethical approach, which still means that you've got to have that human in the loop for now at least anyway, to say, " Right, what is it we're trying to achieve here? Are we being rational? And what are the consequences?" And AI is not yet there, in my humble opinion, certainly not in the mainstream, that can actually make that sort of judgment.

Harry Kemsley: Yeah, I do remember with the conversation with Neil, which was, as I remember back in May, June time last year, there was a big part of what he was saying, not of course just about tradecraft because that was the centre of the episode, but also on this ethics of the intelligence analysis as you just touched upon there. And that tradecraft is an evolving thing of course, and that with the arrival of advanced technologies like AI, machine learning, et cetera, those tactics, those techniques that we're using in intelligence cycle needed to evolve and that evolution was inevitable. I think what I also got from that conversation is that the ability to innovate and evolve is an inevitability has to be done, but the ability to supervise that and ensure that the quality remains high, that's the bit that we're still not very clear about. We're not absolutely certain how we're going to do it. " How do we actually govern what the AI is doing inside the black box?" was a point that was raised at the time. How do you actually drive that machine in the direction you want it to go rather than what it's deciding to go for itself?

Sean Corbett: Yeah, absolutely. And of course that was all brought together by what we call the" super podcast" with three excellent people, Emily, Amy, Clara, and our own Alison as well. Just that they sort of brought together the technical with the ethical and even the empathetic, which as you know was not a word I knew before then and still struggling with, but getting there slowly. But I mean that was a fascinating discussion about how you got to weave them all together.

Harry Kemsley: That's right. When we had those conversations, didn't we, right at the beginning of the year. We had Dr. Claire Yorke in who brought the word" empathy" into your lexicon, Sean, as I recall. And-

Sean Corbett: Yes.

Harry Kemsley: ...we talked about how empathy was a key part of the intelligence process. And then as you say, we then rolled that into the so- called super podcast where we had Emily, we had Claire, we had Amy, and we also had Alison as you said, where we brought together that mixture of ethics, empathy, technology and culture, which was a fascinating discussion. And then as you say, it certainly brought to mind things that I hadn't really considered before. But one of the things that we definitely took away from that conversation is what empathy might actually mean in intelligence terms. If you remember with Claire, we talked about the fact that when you have empathy, you might understand at the state level how that state conceptualizes itself, how it derives meaning, what they talk about when they talk about their past and why that's important to them. It's a way of looking at a situation not just through cultural lenses, but also through a more human understanding of why they might be feeling the way they do, they act the way they do, as difficult as that might be, to really begin to understand what you need to do in response or alongside them. So that empathy piece came through. Now I think at the time you were talking about the fact that empathy was a function of the art of intelligence within that tradecraft piece that of course we talk about so often. And that it's not just something you add on, it's something that's a constant, it's always in the middle of the process. But I think what we took away, Sean, if I remember, is that you and I hadn't really thought about it, it's quite likely others hadn't thought about it in the way Claire had presented it, and that there was a distinct need for us to really think about the process of including some sort of empathic view and ensuring that that mindset we have is taking into consideration other people's perspectives.

Sean Corbett: Yeah, and it is another tool for the analyst in terms of their trade craft is, and maybe we don't want to overblow this because the analyst does try to think in terms-

Harry Kemsley: Yes.

Sean Corbett: ... okay,call it" red- teaming" if you want, but okay, why is a certain individual, Putin, or whoever, is thinking the way he is and therefore acting as he is? So I think we've always done it in a way, but not as much of a conscious way and not weaved in maybe a formal process. But as you say, you can't just say, " Right, tick this box, I've got my assured data, I've got the analytical techniques here. Right now let's apply empathy." It can't work like that. It's far more cognitive.

Harry Kemsley: Yeah. I suppose it's one of those things where we talk about wanting to address potential unconscious bias. If you don't ask yourself that question about why somebody or something is operating the way it is, then you are potentially going to miss the point. You're going to completely react in the wrong way. And I suppose that's the point, isn't it, Sean, that it's not just about the quality of your analysis, it's the breadth and depth of your analysis that you've done and that you've potentially introduced biases by your own culture, education. And that if you're not looking at it through other lenses using the word" empathy," again, you could be missing the point. You could be completely misinterpreting the situation.

Sean Corbett: Yeah, absolutely. And that's actually quite a neat segue into my other really big takeaway, which I know we're going to have to cover, or we will and we are planning to cover again, is misinformation and disinformation.

Harry Kemsley: Yes.

Sean Corbett: We talk about the unconscious bias, and again, maybe it's a bit unfair, but I still go back to the Al Ahli hospital explosion where an individual from the BBC made two- and- two- equals- seven straight away probably because of their unconscious bias, maybe previous experiences where attacks had happened. And I would suggest that having probably been embedded within Gaza for a while, you have that empathy with the organization that you are there, because they're not all murderous terrorists. Immediately that unconscious bias when, " Oh, they've been attacked at that hospital," now very clearly it didn't take as long as experts to go, " There is no way that that was a 500- pound, a thousand pound, 2000- pound bomb that went off there." So the facts very quickly, but the reporting was so instant and so quick that they just got it wrong. Now-

Harry Kemsley: I guess-

Sean Corbett: ...that's-

Harry Kemsley: ...if you, I guess if you-

Sean Corbett: Sorry.

Harry Kemsley: ... livedthe experience of Gaza and you have a variety of ordnance falling from the sky and creating damage, it wouldn't be hard, would it, to take a step-

Sean Corbett: No.

Harry Kemsley: ...from, " An explosion-

Sean Corbett: No.

Harry Kemsley: ...over by that hospital must have been caused by some air- delivered audience or artillery." It wouldn't be hard to make that step. I think the point you're making though is even though you are exposed to that conditioning, even though you might actually be in the middle of all of it reporting from close quarters, you need to step back and just take a breath before you report that, " The Israelis have dropped an air- delivered or artillery- delivered ordnance on a hospital," because that is not what it turned out to have been.

Sean Corbett: Indeed.

Harry Kemsley: Talking about mis and disinformation though, we spoke to Di Cooke, didn't we, back in the early part of last year who, as I recall, we focused largely on deepfakes, the sort of imagery and video technologies that are available. For me, the big takeaway in that conversation was that in the so- called post- truth world where many of us still believe we can see through falsehoods, we can see through the myths and disinformation that's around, the impression I got from Claire was that, sorry, from Di, was that that actually is not the case. We're not very good at all at spotting fakes, particularly in the video and imagery world. She talked about, " Artificial intelligence created deepfakes that was so good, currently the only way of really detecting them reliably as being fakes was to use other AIs." You end up with this war of AI going on in the deepfake arena, which for me, the big takeaway there was we're not as good as we think we are as humans at spotting deepfakes in the video and imagery world.

Sean Corbett: Yeah, that was quite scary actually. I mean, because she's doing her PhD research, she might have actually finished it now actually, but it was obvious that I think it was more than 50% of the deepfakes that were shown to experts were not identified as deepfakes.

Harry Kemsley: Right.

Sean Corbett: And that should worry us. I mean it creates issues for open source intelligence, but it also creates opportunities I think because there's a real key role for open source intelligence there. And I've thought about this a lot more and I know we're going to cover it, we have to because it's prevalent in everything that we do now, misinformation and disinformation. But I guess context is one of the things that you need to help you understand that. So the classic one that everybody talks about in terms of the Pope in a puffer jacket, I mean clearly the Pope is not going to wear a puffer jacket, or is that just my perception? I don't know. But when you see things like that, you go, " That can't be right." When it becomes more difficult, and there's been some ones with Donald Trump, you go, " Actually that could be realistic, but is it deepfake? Is it real?" There's an example which I don't know if we covered before, but there was a chap in Germany, I think it was, who took a load of mobile phones as an experiment. It put them in a little handcart, put them all on, and walked up a street in one of the big cities in Germany. And it may have been Google Maps, it may have been something else, but very quickly they'll be like, " Avoid this place because there's a huge traffic jam." Because the data looked at what was transmitting and the fact that the phones were going really, really slowly and therefore there must be a traffic jam or an accident there. Now-

Harry Kemsley: Excellent.

Sean Corbett: ...I mean just as an experiment, but that could have been quite easily been, well, it was disinformation, to change patterns of behavior.

Harry Kemsley: But then I guess that example, which I hadn't heard before, that actually brings to mind for me, so what do you do about it? I mean, how would you possibly counter a guy walking down the road with a trolleyload of mobile phones, making all the systems think there's a traffic jam? I mean, I don't know. There must be an answer to that, I'm sure. I don't know what it is, but then-

Sean Corbett: Well, the answer is, and it's the first time I've used the word" tradecraft." So when you see something you think, " Oh, right, what's happening there?" There will be CCTV cameras, there will be eyewitnesses-

Harry Kemsley: Right, yeah. Okay, yeah.

Sean Corbett: There will be people there that say, " This is not happening."

Harry Kemsley: Of course, so-

Sean Corbett: And inaudible-

Harry Kemsley: ...straight-

Sean Corbett: ...inaudible, but, you know-

Harry Kemsley: ...straight to multisource-

Sean Corbett: Indeed.

Harry Kemsley: ...triangulate, get the different perspective. Is there a traffic jam there? Yeah, of course. CCTV just pop it up and say, " There's a guy with a trolley, but can't see any traffic," and that would've answered the question. He must have a few telephones in his back pocket. Yeah, interesting. I think of all the topics, and we'll come onto some other ideas for the coming year of all the topics, Sean, we're going to have to come back to this mis and disinformation piece.

Sean Corbett: Yeah.

Harry Kemsley: With Di, we certainly looked at one aspect of it, which was fascinating, albeit worrying. But I do think we should come back and look at that in a slightly broader context because if we believe open source is becoming a very powerful current and possibly predictive intelligence source, well, it is unquestionable that it's got to be reliable and if it's full of misinformation, we need to know how to deal with it. It's got to be one of those things that we've got to get our head around is how do we actually deal with it in the open source environment? I think it's probably true to say that it's going to be less of a problem in the closed information environment because you have more control over what it is that your censors are seeing and doing. I do stress probably less of a problem. I'm sure there's plenty of examples where you could refute that, but in the open source, it's the Wild West. Everybody can say everything and as somebody said a while ago, everybody lies. So, if you've got all the information out there and it's all mostly wrong, how do you weed out the good stuff from the bad stuff, the wheat from the chaff, so to speak?

Sean Corbett: Yeah, I agree with that and this is where responsible open source intelligence does play its part and the tradecraft does come in. I mean one of my real concerns at the moment is the Wild West is, say, of social media, but also mainstream media. Everything is instant soundbites now and they do get it wrong. I'd like to think that some of the responsible mainstream media just gets it wrong, but I'm not sure anymore. And then there's that blurring area between what is deliberate disinformation, what is a unconscious or conscious bias, and what is partial truth? Now, as I said, there is a role to be playing for mainstream media organizations, but I'm not sure they're playing it certainly in the way that they should be. And that's where open source intelligence organizations that can demonstrate their process, that can demonstrate their sources, that do that correlation, that triangulation go, " This is the ground truth, not my truth or your truth, but the truth." And that's really important going forward.

Harry Kemsley: And I do remember from Warren, Warren Strobel from the Wall Street Journal who was with us earlier last year, he talked about, it's inaudible the recording, talked about the fact that 90, 85, 90% of what government intelligence analysts could see in their closed environment, he believed he could see from the open source environment as a responsible journalist.

Sean Corbett: Yeah.

Harry Kemsley: If that's true, and I'm not going to disagree or dispute that, if I take that at face value, then the responsibility of the Wall Street Journal and other media analysts is going to need to be very, very clearly thinking about the impact of what they're saying, trying to be as accurate as possible for fear that they're actually creating more problems than they're solving by the reporting they're doing. Because if we're not careful, they become the biggest source of mis or at worst disinformation, particularly if they're now able to see almost 90%, he was saying, 90, 90- odd percent of what at least an analyst could see inside an agency. Just moving on from that then, just to finish off, because I've just touched on the Wall Street Journal, it's what's reminded me of it, one of the topics that we came up with in our conversation with Randy Nixon from CIA, Bob Ashley, ex- of DIA, of course, was this partnership between the government organizations and the commercial providers. If you remember, we talked about that becoming a necessity, that the government organizations are going to have to start thinking, to be fair, I think they are starting to think about it, it's not like it's a brand new topic for them, but really understanding how do we bring together the classified and the unclassified community to a mutual benefit and better understanding?

Sean Corbett: Yeah, and this is a really big one for me, and I'm sure you are almost as frustrated as I am having seen things from the other end of the telescope where you don't trust industry and thinking, " Well, no, everything has to be done internally," and, " I can develop this," et cetera, et cetera. Well, the fact is that none of us can do it on our own. There are elements both of the government and non- government commercial and non- commercial, that have to work together. And the frustration I has is that while discussions are definitely going on and we've seen them progress, and-

Harry Kemsley: Sure.

Sean Corbett: ...Randy has been great in that respect as have others, but that conversation is not yet at a stage, a level of sophistication where, okay, what are we going to do about it? How does it look? Is there going to be a, for example, a US open source intelligence agency? If so are they're going to do everything themselves? How do you bring the commercial sector in? There are so many elements to that, everything from individual bias and sadly in our world, individuals still really matter. If somebody in an organization that's in the right level just goes, " You know what? I don't want to do this," or, " I don't understand the need for it," then it tends not to happen. So you need for me an enterprise level, but an enterprise level of engagement that says, " Right, we're committed to this." And by the way, never mind all the policies and the funding challenges and everything else that we've got to deal with. We need to make this happen. And let's have meaningful discussions and actually come up with a plan that delivers something. I mean, you and I have both been in discussions with government organizations where you have a great discussion and afterwards all of a sudden nothing happens. And it's not because necessarily because people don't want it to happen. It might be budgetary constraints, it might be just" the tyranny of the inbox" as I call it, that things don't progress, but we're going to have to do this because there's so many bad things happening in the world now. If you look at the threats that were emerging, threats and everything else, there is a huge amount of global insecurity from every perspective that we talked about. We have to do this together and we have to do it in a clever way. I don't think we're there yet.

Harry Kemsley: No. I remember the conversation we had with Robert Cardillo looking principally at the imagery intelligence, geospatial intelligence world, and we touched on this topic about the commercial capabilities of which there are, of course many go back, I don't know, 10, 15 years, how many satellites were owned by commercial organizations that were readily available? Look at it now with your credit card in one hand and imagery in the other hand sort of thing. But he talked about there needs to be a line that you don't cross. He talked about the fact that eventually governments do need to reserve the ability to look after their own national security interests and that that could not be relied upon entirely in the civil and commercial sector. So I think when you blend those two topics, those two points together, you end up with this absolute necessity for a combination, a collaboration between the commercial and the government, but at the same time a recognition that there are some things that the commercial environment is really good at and can do for and with the government, but eventually there will come a time when the government's got to do things for itself for its national security requirements.

Sean Corbett: Yeah, and that's absolutely right. And as you know, I've always been very, very careful to make the point that I don't think that open source intelligence can replace the intelligence community and all the other good stuff, but it does have an important role in the heavy lifting and the data and some of the analysis as well. But it's got to be a true, I use the" I" word for a change, " integration"-

Harry Kemsley: Sure.

Sean Corbett: ...where you're taking, you're reversing the 80/20 whilst since I've said that actually, where 80% of the really good stuff was coming from exquisite sources and 20% was sprinkled on top, if you like, from open source, that is reversing, particularly in the imaging world, as Robert would say-

Harry Kemsley: Certainly, the imaging world.

Sean Corbett: ...where some of the capabilities are really good out there in the commercial world.

Harry Kemsley: Okay. So, 20 episodes, four or five big takeaways. So we've talked about the foundational and importance of open source intelligence becoming increasingly centered around current intelligence potentially even becoming predictive. We've discussed the value and role of artificial intelligence and other advanced technologies, and the balance between the ability for those black boxes to do what they're good at versus the human. We've introduced the idea of the ethical and the empathy aspects of intelligence, which is really just about being much more explicit about those things rather than it being an inbuilt thing, we actually consciously think about it, really important in the mis and disinformation post- truth world. We've recognized, I think in some of those conversations we've had before and in this sum up that there are some things the human can't do as well as we would like to believe we can, that we may have to rely upon the artificial intelligence and other advanced technology. And then this combination, as you just finished off there, talking about the importance of that link between commercial and government agencies, albeit there may be some limits. So I think those are really crisp, very valuable takeaways from the podcast episodes for this year. Sean, what have we not done enough of? What should we be looking at going forward? Rather than talking about the one takeaway for this particular episode, what are the two or three things we should look at for the coming year?

Sean Corbett: We've filed out a couple of them already. We've got to get more into the practical applications for misinformation, disinformation. Absolutely. I think for me, the big one is, and it is being discussed a lot now, is the broad sunlit uplands if you like, of intelligence, and that's predictive analysis.

Harry Kemsley: Predictive, yeah.

Sean Corbett: Being able to forecast, if you call it" indications and warnings," but it's more than that. It's strategically foresight to say, what are the areas in this world that we need to keep an eye on to see, are they going to be emerging problems or are they there, we're just not looking at them?

Harry Kemsley: Yeah.

Sean Corbett: And there are certain hotspots in the world where we may not be as focused as we could be. The Balkans is one area, for example-

Harry Kemsley: Sure. Yeah.

Sean Corbett: ... Libya could implodethis year. There's the ongoing crisis obviously in Gaza, and the opportunity for" contagion," as I call it, around that area. Obviously we've got to keep an eye on China and see what their intent is. So trying to project ahead and maybe even stop crisis before they happen. Now, that is incredibly challenging. Always has been for the intelligence community as a-

Harry Kemsley: Sure.

Sean Corbett: ... whole,but I think it's as applicable and as challenging for the open source community. I've been looking a lot in the terms of the factors recently that led to the Arab Spring or at least triggered it, been quite a lot of that. And actually worryingly, you see a lot of the same factors now-

Harry Kemsley: Sure. Sure.

Sean Corbett: terms of the demographics, in terms of things like drought and things like wheat prices going through the roof. Now it is a different time, so I'm not predicting that there will be another Arab Spring, but there's potential for some areas to see increased unrest, put it that way. So-

Harry Kemsley: Okay.

Sean Corbett: ...and that could have been predicted had we been looking at the right things at the right time. So that sort of thing-

Harry Kemsley: So in predictive analysis, let's get our crystal balls out and stare into them to find out what the year ahead might hold. You mentioned mis and disinformation. I agree. We probably will always come back to that. I suspect that will drag us back into AI, but as it does so, let's start looking at what are the practical applications of AI more broadly? We talked about videos and imagery in the mis and disinformation. We'll broaden that out to something slightly more generic, slightly more wide perspective. Similarly with AI, we talked about the importance of AI in the intelligence cycle. Let's now start looking at that more broadly into the defense environment. What are the applications of AI? And one thing that we sort of touched on with the relationship between government and commercial organizations and then between government non- traditional partners, is the sharing of intelligence. I think we've obliquely touched on intelligence sharing, but I'm sure that recent conflicts, ongoing conflicts in Ukraine and probably other parts of the world are really testing some of those security officers who are driving very hard to say, " No foreign. You can't share this with anybody." Well, actually I do need to share. I do need to share because that non- traditional partner needs to know the following three things because they're part of our coalition. That's genuinely big problem, but it's probably bigger than that. So let's look at that for me, Sean. I think of the ones we talked about, predictive analysis, yes, AI of course, mis and disinformation, it'll be ongoing. But I think as well intelligence sharing, we should try and bring that in as well. What do you think?

Sean Corbett: Yeah, it's a very good point. As you know, it's very close to my heart. And I think most podcasts, we've actually touched on it. But you'll be pleased to know that I do have a guest in mind that is very used to-

Harry Kemsley: Oh, very good.

Sean Corbett: ...inaudible intelligence. So look forward to that and hopefully the coming month actually. But-

Harry Kemsley: And that was-

Sean Corbett: ... it wasvery-

Harry Kemsley: ... very cool, wasn't thatalso your role at the DIA when you were there inaudible?

Sean Corbett: Indeed, indeed.

Harry Kemsley: Right.

Sean Corbett: Yeah, one of my main roles was to try and inculcate increased sharing within the Five Eyes community, which was an interesting challenge in itself. We made headway. It's still a long ways to go.

Harry Kemsley: See, well what we'll do then is I'll light the touch paper between you and the guest and sit back and just listen. All right, well look, I think we've got four or five topics there which we've drawn out from this last year. We've done it over 20 episodes. Let's see what FY'24, the whole year ahead of us brings. I think the four topics we talked about sound like good ones to start with. But I think I should finish by doing two things. First of all, thanking you, Sean, as ever, for the huge amount of effort and work you put in preparing these podcasts, because you do most of heavy lifting on that front and I am very grateful. But secondly, but by no means least, I would really like to thank our many guests we've had this year. We've had some outstanding guests through the year, and if we haven't mentioned you by name and you know who you are, if you listen to this episode, thank you very much. I am very grateful for the effort and time you've put in with these podcasts. They've been great fun to do. I've learned huge amounts. I continue to learn a load. So thank you to you, Sean, thank you to the guests, and of course, thank you to the listeners who've come back and listened to you and I droning on.

Sean Corbett: Indeed. No, it is been my pleasure and I really enjoy these. We get a lot out of them, and hopefully so does the audience. And yes, thanks very much for this, for fantastic guests we've had, university, they've been excellent. But I agree. The very last point is that the listenership, I'm staggered at the breadth and depth of people that actually take the time to listen to our podcasts. And as ever, and I know you're going to say it as well, but if there's anything that you think we should have covered better or haven't covered at all, or would just like to hear about, then do drop us a line and we'll have a look at it.

Harry Kemsley: Absolutely. And I do think we should start going out to some of those locations we've discovered are listening to us, particularly in the Canary Islands. I think we should need to do a podcast recording-

Sean Corbett: Absolutely.

Harry Kemsley: the Canary Islands just for fun. Or if not that, the Caribbean, which are also big listeners of ours. So, without further ado, Sean, thank you for everything we've done together in FY'23 in the year of 2023. Let's see what FY'24, 2024 brings. Thanks Sean.

Sean Corbett: Absolutely. Thanks Harry.

Harry Kemsley: Goodbye.

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.


Over the course of 2023 Harry Kemsley and Sean Corbett hosted more than 20 podcasts exploring topics such as the implications of AI for OSINT, the need for empathy in

decision making, and tradecraft in open-source intelligence. Janes analysts also discussed how OSINT has supported Janes analysis of new and emerging situations such as that in Sudan, Haiti, and Israel. In this episode Harry and Sean look back on the key themes and what they learnt from the discussions with their guests in 2023.

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