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 The World of Intelligence, and a new year. Happy New Year, 2025, Sean.

Sean Corbett: Happy New Year, Harry.

Harry Kemsley: So, what a year last year was.

Sean Corbett: Very busy.

Harry Kemsley: Yeah. 26 podcasts.

Sean Corbett: So, in total, by the end of it, 26 podcasts. Yeah.

Harry Kemsley: 26 podcasts.

Sean Corbett: It was a few more than we anticipated.

Harry Kemsley: It sure was a few more than we planned, and, by the way, for the record, to our listeners, the way we run these podcasts is that I pontificate and I pretend to know what I'm talking about. Sean does all the preparation and all the organization. So, Sean, thank you for all that extra work you've done.

Sean Corbett: No problem.

Harry Kemsley: We have done a great deal of work this year, or I should be honest and say you've done a great deal of work this year getting these podcasts-

Sean Corbett: We.

Harry Kemsley: Thank you. Now just the sheer volume and breadth of challenges that are in the world, it's got us thinking about how OSINT is understood, how it's being used, how it's developing and maturing. We've covered all kinds of topics. And I thought what we might do, Sean, today in the absence of a guest, because everybody else is quite rightly sitting at home drinking a glass of something nice, eating a mince pie, or whatever they choose to do at Christmas and New Year. But let's just go through some of the things that we've done. Now just to get us started, I think of all the topics we've discussed this year, the two that come out loud and clear for me are misinformation/ disinformation.

Sean Corbett: Yup.

Harry Kemsley: And the use of AI. Now there are other things we're going to talk about, but from my perspective, those are the two that really stood out for me. We've spent quite a lot of time this year talking about mis and disinformation, but also about the uses of AI.

Sean Corbett: Absolutely, and that shouldn't be any surprise really, because those are probably two of the most topical elements on not just talking about open source intelligence, but actually society as well.

Harry Kemsley: Yeah.

Sean Corbett: Yeah. AI is pervasive everywhere. There's a debate right now as to have we reached almost like a speed bump in terms of we've got as far as we can as a tool. That's one view of looking at it. To another view, we're nowhere near there yet, and, of course, there's those that think it's the best thing since sliced bread, and those that won't touch it still. And that's reflected some by discussion.

Harry Kemsley: Yeah, and we've had conversations about what is AI. We've had conversations about how AI might be used. We had a conversation just before Christmas and the New Year with Harry from Janes about the pros and cons of large language models, and I think we've really looked at it. I think my takeaway from all those conversations is that as with many tools, if you know how to use the tool effectively, the chances are it will be useful to you and you can use it with some degree of care, but you can definitely get something from it. But used blindly, used without due care and attention, it can really get you into trouble.

Sean Corbett: Yeah. Absolutely, and that was an excellent one to round off the year. Wasn't it? I think as I said at the time that AI is not a panacea, but it's a really useful tool.

Harry Kemsley: Yeah.

Sean Corbett: But you've got to treat it, in many ways, like you would an analyst. I don't mean put it in a box, I didn't mean that, what I really mean is-

Harry Kemsley: inaudible on.

Sean Corbett: Yeah. Exactly. You have to follow the... And it is rather going to be about when you use the word tradecraft.

Harry Kemsley: Yeah.

Sean Corbett: Even if it's inaudible as opposed to an individual analyst called tradecraft.

Harry Kemsley: Yeah. When you think about the words that people like Keith inaudible focus about what it was, and its applications, I think he actually used the phrase, which I quite like, that to not use AI, and advanced analytical tools these days, it would almost be tantamount to negligence in the world of national security, because it is such a powerful tool, if you know what you're doing with it. And then on a similar vein, Martin came in in the future applications space. You're talking about artificial general intelligence, which I know is not a universally accepted concept, and not everyone buys into the idea that we are potentially quite close to it being achieved, but nonetheless. We need to accept that genie is out of the bottle, and we need to start thinking about how we're going to grapple with it and use it. So, that the tool becomes useful to us rather than our master.

Sean Corbett: Absolutely. And I think we'll probably return to that this year just for that very reason that you can sit there in the past, and not engage, but the fact is that AI is going to-

Harry Kemsley: Yeah.

Sean Corbett: ...develop and you've got to embrace it and use it in the way that it's going to be most beneficial.

Harry Kemsley: Yeah. Now moving through AI brings me back to some discussions as I mentioned at the beginning around mis and disinformation, and we talked a long while ago. Didn't we? About the creation of deepfake videos and the fact that AI was creating them-

Sean Corbett: Yeah.

Harry Kemsley: ...and then we talked about how only AI could spot deepfakes, et cetera. Well, I got a very similar feeling talking to the topic through the course of the year in the various podcasts we did, but one of our things that we got very early on last year when we spoke to Amil, Amil Khan, the ex- journalist from Reuters and BBC, was that he mentioned something that has stuck with me ever since, which is that way of detecting that perhaps you are suffering from what some people call the echo chamber, or that you are becoming without knowing about it more and more biased in your views, because of the way you're working with media. And it's something I've actually spent quite a lot of time in my own professional capacity looking at, and looking at how I engage with media. Amil talked about the article heading that you are drawn to, and you find yourself reeling at, is the one that you need to check, because that is probably telling you something about your own echo chamber that you're creating around yourself. So, I think there's a topic, which we might want to play up this year, Sean, about what can I do to help myself ensure I'm getting as close to truth, in, again, what some people are calling the post- truth age?

Sean Corbett: Yeah. Absolutely. And as you know, this is one of the things I worry more about than the potential for nuclear Armageddon. Right?

Harry Kemsley: inaudible podcast you actually thought this was a potential to unravel democracy.

Sean Corbett: Absolutely. Yeah. Yeah. Polarization of society, it really is, and as we mentioned in another means actually that... And I think this is what you're leading to in terms of coming back to, there's a personal responsibility here, in terms of making sure that you're balanced enough in terms of what you absorb and however you absorb it, and, therefore, giving a worldview, so, that actually almost you contribute to not that echo chamber filtered button, but are able to have a balanced conversation. Now how does that lead across to open source intelligence? Well, it's exactly the same.

Harry Kemsley: Yeah.

Sean Corbett: It's that the unconscious bias, which we've got to be so careful about, is that analysts are tempted to use the same sources that they know work pretty well, and they know provide them with a good analysis and filled in some gaps. So, naturally, they're going to go back to those.

Harry Kemsley: Sure.

Sean Corbett: And there's nothing wrong with that. But what you mustn't do inaudible to say, " Well, I'm not going to look at over here for this source," and this is where... And, again, back to the tradecraft piece, you need to, at least, consider, that doesn't mean include, every available source on something. You can absolutely disregard stuff, but you've actually got to go in through each one.

Harry Kemsley: Yeah. I liked the way in our discussions around mis and disinformation, the way of checking yourself includes deliberately go and find an article, or a piece of news for a source that you know is likely to give a slant that you don't normally read.

Sean Corbett: Yeah.

Harry Kemsley: So, if you read this kind of organization's work, and they tend towards the left, or right spectrum, find the alternative, opposite-

Sean Corbett: Absolutely.

Harry Kemsley: ... end ofthe spectrum, and the way it was described to me is if you feel yourself becoming irritated by what you're reading, again, it's a symptom of you becoming perhaps too far down one end of the spectrum than the other in terms of your echo chamber that you're creating around yourself-

Sean Corbett: Indeed.

Harry Kemsley: ...a fascinating discussion. And I think we should pick up on that again this year. We've also spent quite a lot of time, an increasing amount of time looking at specific events that have been going on through the year. Haven't we?

Sean Corbett: Yeah.

Harry Kemsley: We looked at Sudan, Bangladesh, Yemen, of course, what's been going on in Ukraine and Lebanon, and how open source intelligence has been applied to understanding those events. And I think one of the things that's come up for me is that open source may well be now increasingly used as the primary source, at least, initially, and that it might become the first choice rather than the last choice of understanding what's happening in some of these places.

Sean Corbett: Yeah, and that's the debate that's definitely raising within government right now, and we've heard Randy... I think it was Randy Nixon who came inaudible to plan the intelligence of first resorts.

Harry Kemsley: Right.

Sean Corbett: And that's absolutely fine, because of just the sheer volume of data and the all- pervasive nature of it, and the global nature of it. I think we should... And I do always think inaudible that it's not going to completely take over from traditional intelligence, secret intelligence, classified, but it can provide the iceberg, the part of the iceberg underneath the surface. You're always going to need the exquisite stuff that goes, " Okay. I need something more specific. I need to know what a certain person's thinking," or, " I need to know a little bit more about that," that only very expensive, exquisite stuff can-

Harry Kemsley: Sure.

Sean Corbett: ...come against. But going back to the crisis stuff they've been talking about, what's been very clear is that you can get a pretty good picture from open sources, but, of course, as we discussed, there's a real role where maybe two years ago, if we had this discussion, we'd talk about foundation intelligence, inaudible, and all the rest of it. But in terms of the initial triage, right, crikey, something's happened. Where is it? What's happening? Why has it happened? And as we've seen actually with, particularly, the Middle East, but others as well actually, and Sudan you mentioned as well, open source analysts are able to go get a very, very clear idea about what's happening, partly because they're data sources, but, partly, and this applies to all analysts is because they've got the background experience they've been looking-

Harry Kemsley: Right. I think one of the things that's struck me during the course of the year, particularly, in these reviews of specific crisis events is the depth, the breadth, the quality of the analysis has felt very similar to me to the breadth, and depth, and quality of it from the exquisite sources you've just spoken about. In other words, where open source might be disregarded, or previously had been disregarded as a nonsense source, actually it's very clear that with the right tradecraft, the right availability of good sources, you can actually produce a very intelligence, very reasonable understanding of what's happening, and, in fact, in some regards, I think we're getting a better view than we might be getting from some of the more traditional government- based sources, because they have their own agendas. I've found some of the discussions around, for example, what's been happening in events in Iran, for example, have been often more close to what I suspect is actually true than what actually I've read in some of the formal government announcements as we go through all those news sources. So, I think there's a degree of truth, and legitimacy from open source that is becoming clearer, and clearer to me.

Sean Corbett: That's a really interesting point, and one that's being discussed at the moment is that to what extent is the intelligence community becoming politicized? Not necessarily to a certain fault, but, of course, it's all about who is asking the questions, and what questions are being asked.

Harry Kemsley: Sure.

Sean Corbett: But I think, and one area I'd like to get back to that we covered briefly, but I don't think we've done anything really in- depth about is culture. The culture of the organization from which you come.

Harry Kemsley: Yeah.

Sean Corbett: And that includes everything from diversity to exactly as you've said in terms of the sources that you're using, the perceived wisdom, " We do things this way, because we've always done them this way." And the open source creation community is not immune to that at all either, but it's how you optimize the culture to come up with the best assessment analysis, which I suspect there's some PhD thesis in there. It's probably already been done, but I would like to come back to the cultural issue at some state.

Harry Kemsley: Yeah. We've talked about culture in the past in terms of resistance to open source, and we've also talked about it around AI in terms of, I think you said a few minutes ago, the two camps of AI. It's nonsense, it's got to its highest degree of capability, or it's got a long way to go. It's another area for us to discuss in the next year I'm sure. One of the things that I recall from our discussions around particular scenarios that we looked at in terms of inaudible is we also looked at some predictive stuff. Didn't we? We talked about-

Sean Corbett: Yeah.

Harry Kemsley: ...how we might use the open source arena to look at what might happen in the future, and as I recall we talked about... I think you called it the high north-

Sean Corbett: Yes.

Harry Kemsley: ...the Arctic region.

Sean Corbett: Yes.

Harry Kemsley: And looking at open source for the power it can bring in terms of predictive understanding.

Sean Corbett: Absolutely. And there's a piece of what's going on that we will do a podcast on actually with a couple of the analysts here about the likelihood of China invading Taiwan.

Harry Kemsley: Taiwan. Yeah.

Sean Corbett: So, of note, is INW indicators and warnings, and actually what's been really interesting is seeing Janes' produced a fairly sophisticated model that will have a look at that. Forecasting, or predicting it is the nirvana for the intelligence world. The amount of times that I have had a senior guy, tongue- in- cheek maybe, or maybe not, say to me-

Harry Kemsley: Sure.

Sean Corbett: ... "So, when I write the paper today now it's going to be what's going to happen"-

Harry Kemsley: Yeah. What's going to happen?

Sean Corbett: ...which is the... And this is where, again, AI comes in, because if you can do trend analysis and model analysis to say, like, " If the same conditions happen as happened here, this is the most likely outcome," then we've got a better chance. But we will... I think we should... Going back to that, we did quite well, if you remember, on the Venezuela elections.

Harry Kemsley: Right.

Sean Corbett: We put ourselves out there, and the day before, I think it was the... Yeah. It was a couple of days before when, " This is what we think is going to happen," based on a lot of analytical rigor, and actually we were pretty much 100% right.

Harry Kemsley: Yeah. Spot- on. I think the questions that we know all organizations are trying to ask are what's happened? And why has it happened? And then what might happen?

Sean Corbett: Yeah.

Harry Kemsley: Leads very nicely onto, " So, how do I keep good things happening and how do I stop bad things happening? And then what can I learn from it?" Well, those five questions for me are the fundamental questions that any analyst is trying to answer-

Sean Corbett: Yeah.

Harry Kemsley: ...in one way, or the other. So, in terms of looking ahead, we've talked this year about some of the really, really important insights you can get from open source around crisis events. We didn't ... I forgot one actually. We talked about some of the more difficult targets like-

Sean Corbett: Oh, yes. Yes.

Harry Kemsley: ...the ability to get behind the scenes. That was a fascinating conversation we had earlier this year as well.

Sean Corbett: Yeah. It was, and I talk about North Korea as the hardest target. I spent a lot of time looking at it, and there's so little intelligence out there, although, we may get some more out with... Every cloud has a silver lining. If we get reflections from the North Koreans on the battlefield, hopefully, off the battlefield, then we might start to get that, but it's such a closed society it's hard. But the really great thing about that was two very clever analysts who really know their stuff. We're looking at a lot of the diplomatic activity.

Harry Kemsley: Yeah. Yeah.

Sean Corbett: What was said, and-

Harry Kemsley: What wasn't said.

Sean Corbett: ...who was the audience? And what wasn't said. Exactly.

Harry Kemsley: Yeah.

Sean Corbett: And, actually, everything that they predicted is coming to us now. So, at the strategic level, I think there's a lot to be said for that level of analysis. And then how effective the North Korean troops are, we may know in due course, et cetera, but-

Harry Kemsley: Sure.

Sean Corbett: ...if you said to me... You'll see figures, things like, " North Korean troops, how many are there actually?" Active troops in North Korea. And there's anything from 500, 000 to a million and a half plus. Because we just don't know. You can have the best imagery in the world, and you're still probably not going to know that until the society is open. But that was a great example of where you can still use commercially, and publicly available sources. As I said, diplomatic meetings, all the rest of it, to come out, at least, with the intent actually, it's one of the times where you can come up with the intent as much as the capability.

Harry Kemsley: Yeah. I agree.

Sean Corbett: More so inaudible around.

Harry Kemsley: What I'd just speak around about that conversation with North Korea in mind is that you can't just look at a snapshot of a specific diplomatic event. You've got to have that long- term look. You've got to understand what looks normal. And equally, be able to understand the language being used, because one of the things Christina was saying very clearly was, " You've got to understand that they use what language in a very deliberate way. They're trying to say something. They're trying to communicate something," and it's not just a random set of voice they're using. They're being very specific. Equally, who's saying it? How senior they are, how close they will be to the power, the sacred power-

Sean Corbett: And that's a great example of both of the analysts we had on are very steeped in that subject. They are experts.

Harry Kemsley: Indeed.

Sean Corbett: And really know their stuff.

Harry Kemsley: Yeah. All right. So, what else could we bring into this conversation looking forward? I think there are probably three things that we really want to pick up on. You talked about culture, I think we're going to talk about AI misinformation/ disinformation again. I think there are a number of things we could do. We've got a conversation that will be had with Claire. Claire-

Sean Corbett: Yeah.

Harry Kemsley: inaudible.

Sean Corbett: Yeah.

Harry Kemsley: We've not done enough on that, because that's well into-

Sean Corbett: I'm not sure if that's going to have been published yet. It's probably the next one to be published actually. Fascinating discussion. Again, back to AI. Can it detect nuances in the way languages are used?

Harry Kemsley: Yeah.

Sean Corbett: And I remember one of your original vignettes, you talk about tornado. Is that a weather phenomenon, or is it a type of aircraft?

Harry Kemsley: Yeah.

Sean Corbett: But it's a lot more than that, and I do recommend the podcast when it comes out, because just the turn of phrase-

Harry Kemsley: Yeah.

Sean Corbett: ...which if you translate it means absolutely nothing... I won't spoil the effect of it, but, yeah, language is important, and we probably will-

Harry Kemsley: inaudible podcast recorded is that it's one of a couple now where we've been asked questions-

Sean Corbett: Yes. That's true.

Harry Kemsley: ...inaudible to bring that up.

Sean Corbett: That's absolutely true.

Harry Kemsley: I also realize as I'm beginning to bring this conversation to a close, the one thing we didn't talk about was that podcast we had about whether OSINT needs to exist as inaudible move on.

Sean Corbett: So, I was going to come to that actually, because the thread throughout all of these, of course, is as we want to is to expose open source intelligence to the masses. What does it mean? But also the journey it's on, and I think the word is journey, to the extent that... A little bit tongue- in- cheek, but to distinguish academics that postulate that there's no such thing as OSINT.

Harry Kemsley: Yeah.

Sean Corbett: What they meant actually, maybe what they meant, and people will deduce their own from it is that we've got to stop thinking of it as its own discipline, because it's multiple disciplines, if you like. So, inaudible human to all the rest of it. But in the open source domain. So, does that make any difference? And there's always a discussion to that, but we won't go into too much. But I think that the discussion needs to... And we're ahead of the old acting it now. So, what's next? What's next? What are we looking at?

Harry Kemsley: Yeah.

Sean Corbett: Meanwhile, the poor old intelligence community is saying, " Well, I haven't got funding to actually set up this particular open source intelligence. I haven't got enough." So, we need to be careful. But this is where I like to think that we're helping the debate move on, because I think what we need to have a look at next is beyond open source intelligence. I've been thinking about this quite a lot actually, because all of the tradecraft, so, the analytical standards, the way we do business in terms of making sure we've got the right sources and we're not overweighing them, and all the rest of it. It applies to pretty much everything where you need to have a good understanding of what's happening, and why things are happening. And so, that applies way, way beyond those open source intelligence. And I think we should probably start to look a little bit more about what we know as the non- traditional threats.

Harry Kemsley: Yeah.

Sean Corbett: So, climate change, apply the same rigor to the analysis of climate change, that will be an interesting one-

Harry Kemsley: Yeah.

Sean Corbett: ... andmigration and all the rest of it, because that will inform... And there's some debate over there, which is not particularly well- informed... Now, obviously, there are analysts that do that rigor, but it doesn't happen always. And so, I think we should do a little bit of the non- traditional stuff, but also in a context that it's not about open source intelligence. It might be about where do I invest money?

Harry Kemsley: Yeah.

Sean Corbett: Because what's the best environment to do that in? And it could be from a threat perspective, but more likely it's from a, " Okay. Do we have the people who will invest in it? What's the economic climate like?" Et cetera, et cetera. So, I think taking it beyond open source intelligence, I use that rather than most inaudible open source intelligence. And, of course, if you are sitting in NATO right now, you're as interested in the capabilities of your partners and allies, so, that is definitely not intelligence, as you are inaudible you've probably been studying forever-

Harry Kemsley: Yeah.

Sean Corbett: ...but if you've got an interesting partner that wants to come and say, " Look, we like what you did with part of it," you need to be able to assess, " Okay. Excellent. What can you bring to the party, and what do we need to increase your interoperability? And bring you onboard." So, that is just as valid as-

Harry Kemsley: I wonder whether the beyond OSINT is really saying that to have described something as inaudible open source intelligence has been useful, because it's focused their minds.

Sean Corbett: Yeah.

Harry Kemsley: It's got people to think and talk about it in a way that otherwise might have just been not done. For example, although, I hear it a great deal less these days, I still hear too frequently a statement that shows that inaudible statement thinks that open source intelligence means access to social media that has been summarized.

Sean Corbett: Correct.

Harry Kemsley: And a sentiment analysis is done on it.

Sean Corbett: Yes.

Harry Kemsley: And that somehow that is open source intelligence. It is absolutely part of-

Sean Corbett: An element of.

Harry Kemsley: ... butit is not all that could be done from the open source. And I hope that what we've done this last year is show all the things that we've applied it to are all part of open source intelligence, and that, therefore, talking about open source intelligence as, in quotes, " A discreet capability has been useful, if only, to focus the mind." Now whether that is actually on a continuum, so, that if you're doing, I don't know, imagery analysis and you work for the national geospatial agency in the U.S., you're doing it with highly classified sources, but you're still doing geospatial intelligence analysis, the same kind of analysis-

Sean Corbett: Indeed.

Harry Kemsley: ...actually these days with other subtle language you can buy with a credit card.

Sean Corbett: Yeah.

Harry Kemsley: An average person, in theory, could do.

Sean Corbett: Absolutely.

Harry Kemsley: And I think that's the important distinction is that it's been helpful to address the matter that isn't the capabilities are available in the open source environment, and now it's focused the mind. Please be careful, we don't know the status. Because the creation of good open source intelligence centers is potentially creating a series of unnecessary complications inaudible.

Sean Corbett: Yeah.

Harry Kemsley: For sure. All right. So, let's now put our marker down for the year coming. Now from my perspective, I already know what I think is my most important takeaway from 2024, and carrying into this year 2025. So, I'm going to give you a moment to think about what your most important takeaway is, because, for me, that inaudible second is where I'd like to focus as much attention as we can afford it. And that is that ... And actually, I think you used the phrase earlier, Sean, I think all of us should have, particularly, people listening to this podcast, who we could assume care about matters of intelligence and truth and getting to the bottom of things. So, if I assume the listener is the kind of person that cares about it, then let me offer you this, I think there is a personal responsibility for everybody to try and be as" data literate" as possible.

Sean Corbett: Absolutely.

Harry Kemsley: In the modern day in order to deal with the ridiculous conspiracy theories that seem to bounce around in the information observed, the mis and disinformation that we suffer from, and, equally, our ability to look at a piece of news, or a bit of information that's come to us, and recognize its potential flaws. The example that we spoke about recently, Sean, before this podcast about how frequently statistics are used to demonstrate a case, and you find that by looking at your statistics, when you actually get underneath the hood-

Sean Corbett: Yeah.

Harry Kemsley: ...it's not just about the fact that it is misleading. It's that it is conveying a very, very different outcome to the one that's actually the case, and that that is dangerous, in my opinion, and it's swaying opinion. And that if it is swaying opinion, and you are letting it sway your opinions, and you become more and more polarized, that's not healthy. So, for me, my big takeaway from last year that I'm carrying into this year is all about the personal responsibility of data literacy to be able to deal with these kinds of things. So, Sean, what's your big takeaway from last year?

Sean Corbett: Actually, just adding onto that I think, because it's very, very symbiotic, there's going to be events that we haven't even thought of next year, this year. Black swans. There's going to be some really difficult inflection points strategically I think, and, of course, ethics is going to become even more front and center. And we've seen that already with ICC and all the rest of it, as I say. But there's several underlying themes of which you've just mentioned a few there, but I think that from our perspective and from an analytics perspective, stay inquisitive, stay objective, and whatever the subject is, is use analytical rigor.

Harry Kemsley: You almost said tradecraft. You almost said-

Sean Corbett: Indeed.

Harry Kemsley: All right. Well, look, thanks for listening. Happy New Year to you all, and let's see what 2025 brings. Thanks, Sean.

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