Clionadh Raleigh is not only an accomplished academic, she is also founder of ACLED - delivering the most comprehensive and timely datasets on armed conflict, registering over 300,000 events annually.

Tune in and hear why Clionadh couldn’t care much about AI and why it triggers Lars Peter - who spent the past five weeks in Bob-the-AI-Builder mode (check episodes 84 and 87). You will also hear why Clionadh is considering sending her husband to an ISIS controlled area to study climate adaptation!

On a more serious note, the conversation debunks the humanitarian business myths on climate change and conflict. And explores how thinking about yourself as “the good guys” is harmful.

Check out ACLEDs great work here and enjoy the conversation !!!

00:55 Lars Peter Nissen

This weeks guest on Trumanitarian is an entrepreneur. But she is best known as an academic, and a really good one. Few people understand the anatomy of conflict better than she does. But the way she does her business is very entrepreneurial. More than 20 years ago, Clinaudh Raleigh could not find the data she needed for her PHD and so she then began collecting it. That is in itself not so unusual. But the more unusual is that she kept on going collecting data and today that data collection has become the armed conflict location and event data project, ACLED, the most timely and granular dataset we have on armed conflict. ACLED registers more than 300,000 armed conflicts every year and employs more than 250 people as data collectors and analysts. We talk about this and much more in this fast-paced, highly qualified and not at least fun conversation. Once you have listened to it, check out ACLED and all of their amazing work and once you have done that, let us know what you think on email info@trumanitarian.org or LinkedIN. Review Trumanitarian wherever you get your podcast and make some noise on social media. As always, most importantly, enjoy the conversation.

Clionadh Raleigh, welcome to Trumanitarian.

2:22 Clionadh Raleigh

Thank you very much. It's a pleasure to be here.

2:23 58 Lars Peter Nissen

I have been looking forward to this because you are the founder and the CEO of ACLED. And of course, in ACAPS, you are one of our favorites. We really like you guys. You produce fantastic data. But there might be listeners who are not familiar with Aclid. So could you just take us through what is the origin story of ACLED?

2:45 Clionadh Raleigh

Sure. I'm happy to. I also am a big fan of Trumanitarian, so I'm happy to be here and chatting with you, Lars. So, ACLED started when I had to write my PhD and I realized that there were no data. Everything at the… so I was doing a PhD on conflict patterns across African States and all of the data when I started, many, many years ago, were on the country year level. But all of the things I had read about the conflicts implied that...There was a lot of subnational variation and that really those conflicts often differed from each other. There could be several conflicts occurring in one country. There could be conflicts that have a very distinct spatial signature. And there was absolutely no way to investigate this. And I was doing a PhD in political geography. So it was really vital that there was some sort of subnational variation that I could pick up on. And I had gone to a talk where somebody had mentioned that the U S military had collected quite a lot of information on where they bombed in Vietnam and the variation of those patterns. And I thought, Oh, well, I'll just do the same, but I'll just do it for Africa. All of Africa for, I don't know…at that point, I think it was, it would have been about 35 years. And so I had a research position at Prio at the time and the year looked pretty free so I just started to collect event data on what happened across African countries which in fairness were only six countries I had chosen in Central Africa during that period and I set about reading everything I could and recording systematically as possible what had happened and from there you know ACLED grew. I did that all year and then subsequently for the rest of my life.

4:48 Lars Peter Nissen

And which year was this?

4:50 Clionadh Raleigh

6:09 Lars Peter Nissen

And so today, how many staff do you have in ACLED?

6:13 Clionadh Raleigh

So, ACLEDs size varies between about 240 people to 250. And there has been, of course, like real increases, you know, thresholds that we reached, and then we decide to do something else and it adds a whole extra layer to it. But once we decided to go global, I think we had about 200 people working at ACLED. And then from there, we've really kind of refined the whole process of running kind of a global organization. And we're, I would say about 250 is typically how large we are.

6:51 Lars Peter Nissen

That's really impressive. And I think it's probably fair to say that you are the leading source of data on conflicts at the sub-national level. Nobody else does what you do.

7:00 Clionadh Raleigh

That's right. Which is quite a goal, given that, you know, it kind of started with me by myself with a notebook. But I will say that you know, the team itself is really phenomenal. And what we've tried to do is to, is to make sure that we get extremely locally relevant information and we code it systematically enough so that anybody can use it. Because of course, if you are somebody who's, you know, if you're a mayor in Mexico or you're a Cabela leader in Ethiopia, and you want to know what's happening within your larger region, you're able to access that information. And it's as clean and as useful to you as somebody who might be sitting in, you know, the FFO in Berlin and what they needed for it. That's the point is that it's quite democratic in its distribution. And hopefully of course, then easily able to use.

7:56 Lars Peter Nissen

So, Clionadh, for close to 25 years, you have been collecting event data on conflict. And today you, I mean, it's impressive. You have 250 people around the world doing this. How much data do you actually collect?

8:09 Clionadh Raleigh

So we collect or we release on average about close to about 29,000 events per month and then between 6,000 to 7,000 events a week. We have a kind of an aggressive also supplemental or back coding strategy where we, for example, will find a new partner, they have collected information, it goes through a really rigorous set of tests before we include it. But as an example of this, I think we've got something like 16,000 events from Myanmar that are being checked before they're released, including all the events we've also released from Myanmar. So that number can fluctuate quite a bit, but suffice to say we're quite busy.

8:57 Lars Peter Nissen

That's around, if I calculate correctly, around 350,000 events a year. That is massive.

9:03 Clionadh Raleigh

Yeah, it's quite a lot. It's quite a lot. And of course, like the data itself have now grown to a size that I think it can actually be quite difficult for people to use unless they're able to, unless they have quite a number of skills and are quite data literate. And so our response to this, because it's, you know, quite serious barrier, I think, to entry, has been to try to create several tools where people can log on, well, they actually can just go to the website and say, I want to know more about what's happening in Cameroon, or I want to know more about cartels. And you can filter and look at and graph the data directly without having to download anything by using the, you know, several tools now that are available, including the Explorer and the Trendfinder and then the Calculator for Exposure.

10:02 Lars Peter Nissen

Okay. So you have sort of infographics that makes it really easy to sort of explore your data or, oh, infographic got you. I saw that. Oh, what happened there?

10:15 Clionadh Raleigh

Yeah. Okay. I think that there's been a dashboardification of research and I absolutely hate it.

10:23 Lars Peter Nissen

Yes. Okay. Yes. I, so I totally agree with you, right? It's incredible the amount of garbage that is being produced in that space. I really. But then how are you different? Let me ask you. Yeah, I'm sure you're different. How?

10:37 Clionadh Raleigh

Because, well, we are different. Yes. Thank you. So, so these are, these are effectively like fancy calculators. So the infographics are things we certainly do, and I actually find them super useful, especially for our observatories. If you want to look at, if you want to look at, for example, how the Huthies are disturbing trading routes in Yemen. We do have an infographic for that's updated pretty consistently, et cetera. By the calculators, what I mean is that if you and I were having a debate about how violent one country is over another, it's very quickly answered rather than us downloading the over 2 million events and then doing, putting it into a program and figuring it out. You and I can sit down and say, well, all right, I'm defining violence this way and so this country's trajectory over the last five years has been this and you'll say, but this trajectory is this under this different metric.

11:30 Lars Peter Nissen

So it's a very, not very technical way of exploring what your data actually says…

11:37 Clionadh Raleigh

Yeah, I'm very old school, right? Like I like a calculator, I like a simple and informative graph, I don't like a lot of bells and whistles. And so the tools are trying to appeal to somebody who might be in...Karachi and needs an answer immediately. You know?

11:59 Lars Peter Nissen

I think I understand that. And maybe you could think about adding just one bell and one whistle to your thing and you could call it an infographic and it would still be useful.

12:07 Clionadh Raleigh

Well, the infographic is to me is a static picture of something, right? And you can produce those and then we also have dashboards, right? But I just think that a dashboard requires us to make a huge amount of choices for you about what we think you need to see. A calculator presents you with a few very direct choices in order for you to get the answer you want. That's how I would distinguish it.

12:35 Lars Peter Nissen

So you get reliability without sort of shaping the user's experience too much. Is that what you're saying?

12:43 Clionadh Raleigh

I think that one of the issues I have with conflict analysis more broadly is that we have shaped what we mean by conflict to make it...often extremely easy for us to collect information rather than it being always a valid representation of what's happening on the ground. And ACLED was designed specifically to move away from that. So we don't have a fatality criterion as an entry point for an event. We don't define conflict simply in terms of who's fighting the government. There was a lot of ways where I felt like as an entire community, whether it's practitioners or it's policymakers or it's the media or academics, we needed to get out of this sense of, we only define conflict in an incredibly narrow way. And so therefore the violence that's happening to most people is unseen. That's, that's very much what ACLED set out to do.

13:46 Lars Peter Nissen

Okay. So two officers from the army get drunk in a bar and start shooting each other. Is that an event?

13:50 Clionadh Raleigh

No, that's a personal animosity. However, I will say if somebody comes in and shoots them because they are members of the army, that's an event.

13:56 Lars Peter Nissen

How would you describe an event in layman's terms? What's an event?

14:03 Clionadh Raleigh

Yeah, that's a great question. So, it's technical aspects are that it's an act of violence or a demonstration that happens on a date in a place between a political entity - either two political entities or one and civilians. And the reason being is that if I were to describe it only as two armed, organized groups, then I would miss, of course, the fact that armed, organized groups spend about 35% of their time attacking people who are not armed or organized. So the distinction is this, right? If I'm the RSF and I go into Al-Fashr and I fight SAF for… and one of my battalions fights South and another goes to burn the outer regions of the city. Those are two different events.

14:51 Lars Peter Nissen

That's very granular.

14:53 Clionadh Raleigh

It is extremely granular. We have, we have discussions that don't center on the motive of the individual or the motives of the group, but rather the distinctions between did two things involving the same group happen concurrently or did they happen subsequently? Or did they happen in a way where it's being reported that it was concurrent, but it really was over the course of two evenings or whatever the case. So those are a lot of discussions happen on that granular level. Again, to make sure that every event is coated with the same amount of rigour, which means that if I was in, oh, I don't know, like Moldova, I can be confident that these events that I'm seeing as happening in my own country were coded with the exact same type of systematic approach than in Montenegro. And it's very important that that's the case. In fact, I mean, I don't mean to speak for you, Lars, so I won't, but I'll say this. One of the reasons I think that ACLED is so used by other organizations who require, let's say a conflict feed, is that... It's incredibly systematically and cleanly coded. So there isn't a lot of mistakes. In fact, I hope there's no mistakes that people have with the API or anything else that allows them to feed into other data that they can use.

16:22 Lars Peter Nissen

Yeah. And I think, I mean, you, you obviously come out of an academic background. And I think that somehow shows in the way you do things. We in, in ACAPS, we of course come out of, uh, I think a somewhat more operational background or humanitarian background. I think we don't see ourselves as a think tank or as academics, but as a do tank that tries to influence the operational decision making. And rather than the perfection you talk about, we probably think more about good enough, what is good enough to actually make the difference, right? And of course, then the question is, when is that too much, right? When is the quality issues with the actual data a real problem? But I think it's interesting and it's clear, I'd like to really explore that value that comes from the approach you have. Right? And I think what we do is complementary in some ways, right?

17:17 Clionadh Raleigh

Yeah, very much. I will say that, of course, like you're right that we've spent an awful lot of time kind of generating and keeping a very robust methodology for ACLED. I think that it can become too complex for people to understand some of the choices we've made. And that's a real problem for me. And I think increasingly it's for others because the entire point is to put data in people's hands that they can use. And so we are going through a simplification process, I think soon.

17:51 Lars Peter Nissen

So, I mean, you collect 350,000 events a year. You have 250 people. It's not free to run ACLED. What's the business model here?

18:00 Clionadh Raleigh

That has also been an evolution, I would say. So you're right in the sense that we came out of an academic background and for quite a while, ACLED was supported by my academic grants, with then smaller grants, especially for particular regions people wanted us to start exploring and moving into. And then it became a real problem of scale. So we needed to have a certain scale. Lars, I'm sure you're familiar with this… we needed to have a certain scale and a certain level of distribution and coverage, but there was no funding model for that type of activity to continue. There's often a lot of funding to generate a data set, but not to continue a data set.

18:54 Lars Peter Nissen

Yeah, that sounds extremely familiar.

18:59 Clionadh Raleigh

And so then the issue becomes, are you vital enough to funding agencies that they will work to create something new, to create, to fund sustaining something. And what we did there was working with three governments actually at the time, it was the US, the Dutch and the Germans, who started working with the UN in order to create some sort of a data sustaining project, which ended up being Craft which is a really wonderful effort to make sure that data that people rely on can continue to operate. But I will say that...

19:43 Lars Peter Nissen

I think we should say that Kraft essentially is a pooled fund where several of the donors throw money into one big pot. It's managed by the UN and then they dish it out to people like ACLED and ACAPS. And we love you Craft even though you didn't fund us in the first round.

19:59 Clionadh Raleigh

So I love Kraft and they did fund us. So we're very happy about it.

20:07 Lars Peter Nissen

There's a message there.

20:10 Clionadh Raleigh

Yes, but I will say that like Craft is effectively functioning as the foundation of what we're able to do. And then we can, of course, add other money to continue because it doesn't fund our entire operation. It's in its very early stages. So there was lots of iterations about how it would exactly work. But we came to a point whereby we had a few different problems that we were trying to solve with our model. And the problems were this. We were releasing data weekly, and we found that a lot of businesses were downloading the data and then reselling it as their own with very, very superficial changes. And that became an issue, of course, because they could… they could make quite a large profit out of our, what looks like free work. So we had a business problem whereby people were downloading the data and using it to sell. We had a similar problem in that other organizations were downloading the data, making again, often quite small changes, and then saying that it was a data set for them to be funded to continue to create. And then we had the problem of not knowing who was downloading the data and so not being able to make sure that we had, we have an ethical obligation to make sure that the data is not being misused. And so we came up with two complimentary systems. One was that everybody had to register to download the data. And that seems like a pretty normal thing to do. I registered to read the newspaper, so there didn't seem any issue to be able to register to use the data. And once registering, depending on, but depending on where you came from, you were then granted certain access levels to the data. So for example, if you were from the UN, you have total free unlimited use of all data, all raw data, as you wish. If you were from academia you got six full downloads a year, so you can download the data set six times. And if you want more, you just ask us for more. And the reason that that happened is not because we wanted to restrict movement, but we did need to make sure that again, it's not being repurposed and redesigned for alternative purposes without the citation and of course the legacy of coming from ACLED. Conflict analysis in academia is a bit like the Wild West. So I think that putting some structure in place and of course, putting some structure in place to the data creation itself, as well as who and how we use it is a very typical process. So that allowed us to make sure that people who were selling the data had to acknowledge that that's what they were doing. And so we have a very different relationship with them than we might have to, for example, a user from the UN.

23:17 Lars Peter Nissen

And have you been able to monetize your data? I mean, if those guys are selling it, they should pay you, no?

23:23 Clionadh Raleigh

Yeah, we do monetize some of it again for private use, for commercial use. So, you know, Meta uses the data and other similar companies would use it.

23:37 Lars Peter Nissen

And do you have a policy of like an arms producer cannot use our data to predict where the market will be good next year?

23:44 Clionadh Raleigh

Yeah, we do. We have an ethics committee that where we review all of our customers and all of our clients and their use cases. And we certainly, you know… is it Blackwater? Blackwater?

23:57 Lars Peter Nissen

Yeah, Blackwater. I think they rebranded, I think.

23:59 Clionadh Raleigh

Yeah. They rebrand quite frequently, I think, but we're on to them.

24:03 Lars Peter Nissen

Okay, but that's so if you don't mind, so oil companies, would that be okay?

24:09 Clionadh Raleigh

Yeah, I think that's a legitimate business. I mean, I'm not here to, I'm not here to judge the specificity of whether or not I think that their business model is all right. It's whether or not they're using our data for nefarious purposes. You know, whether or not there's sanctions against people or organizations, et cetera, is really our concern.

24:28 Lars Peter Nissen

Okay. So we're really talking about the bad, bad guys, not like big tobacco. That would be okay.

24:34 Clionadh Raleigh

I mean, I don't know why they would need conflict data, but if they do, people smoke and I'm not here to judge them for it.

24:40 Lars Peter Nissen

Okay. No, that's very clear. It's very clear. And Meta, I mean, there has been, you know, there's been some issues where at least it's been alleged that some of Meta's platforms have been used to spread disinformation and create conflict.

24:57 Clionadh Raleigh

Yeah, that's right. And I think that Meta, in my understanding of it, and we've had quite a number of calls with the people who work there, is that what they're trying to do is to understand where they would need higher levels of oversight of what's happening within their platforms when they know that there is a problematic hotspot, et cetera, happening. And I think that that is their due diligence to know that, especially if they believe that their platform is then being used to escalate situations.

25:28 Lars Peter Nissen

Do you find it easy, really difficult to figure out who to give the data to? Is it sort of like you see it and go, oh yeah, that's clear? And that's a no, and this is a yes?

25:37 Clionadh Raleigh

Yeah. I...I suppose I find it quite clear in that I am not here to judge somebody's business model. If it's not necessarily something I engage in or something I support, that doesn't mean that it's wrong. What we're trying to stop is illegality on our part, but also on their part using our data. So it's not judgment. It's, you know, it's illegal behavior. Are you horrified by my response, Lars? Is that what the silence is about?

26:10 Lars Peter Nissen

You have a very clear policy, a very clear answer. I think that's very straight. The horrified doesn't even answer the equation.

26:16

Well, I'll horrify you a bit more if you wish. I'm starting to think that quite a lot of our efforts, not the humanitarian, but I think some of the development efforts and especially ignoring the investment opportunities that might be available for some countries that are receiving huge amounts of humanitarian and development funds is perpetuating violence. I think that the development sometimes perpetuates the violence and the investment might work against it.

26:48 Lars Peter Nissen

I don't know about that. I don't know a yes or no, but what I do know is if we think of ourself as the good guys, if we call the people who use our services beneficiaries, then we have already defined that what we're doing is beneficial. And I think that is so dangerous to assume that we are the good guys. I don't think we consciously are the bad guys, but we have to be careful. I think we're just the guys and the gals.

27:18 Clionadh Raleigh

Yes. I mean, I think we're providing a service and we should not deify ourselves.

27:26 Lars Peter Nissen

Deify is the word I was looking for. Thank you. All right. I'm glad we cleared that up. No, on a serious note, I really like your approach because I think it is so disingenuous, a lot of the discourse around our self-understanding, the way we talk about the humanitarian sector. That's why I more or less consistently call it a industry, because I think we need to talk about the underpinning, we need to talk about the incentives and our own business model that underpins what we do, because that clearly shapes the way we operate.

27:58 Clionadh Raleigh

Yeah, and I suppose like, I think that sometimes the deification allows us to be ineffective, inefficient, not necessarily, not necessarily, I would say, impactful in the way that we should be, because we're already good. So we don't actually have to be good at what we do.

28:16 Lars Peter Nissen

It's a savior complex and you don't fuck with Jesus.

28:20 Clionadh Raleigh

Yeah. Well, some people did in fairness, but yeah, I get it. Yteah. But I think that it's related in my sense to this view of conflict as a breakdown. If you've decided that conflict is due to a breakdown in governance or the economy or the right way to do societal, let's say, engineering, then it gives you an enormous amount of moral license to dictate the terms of the political economic system that you deem sufficiently peaceful for people to have, even when those very changes in political and economic, let's say functioning within a country is exactly why there's conflict. I think that people remain totally blind to the way in which we have, through our engagements of political engineering most specifically, generated violence.

29:22 Lars Peter Nissen

I agree with you, right? And I think there's such a strong parallel between what you're saying and then the way we sort of shape the humanitarian narrative. And for me, the fundamental problem is that it's sort of monopolized, it's not questioned. And the people who are shaping the narrative are the same people who pretend to be the ones solving the problem they've constructed. That's not very healthy. It's called a self-licking ice cream cone.

29:50 Clionadh Raleigh

Yeah, exactly, yeah. Well, they don't have to live with their diktats, you know? And that's the real problem here is I think that we've, this is associated with that, but I always think that when I look at a decision made by a policymaker for, let's say, one of the countries that we might be most active in, it seems like it's entirely based on the internal politics back home, wherever that is, whether it's DC or London or Geneva or wherever else, it's very, it almost, or it's very rarely has anything to do with the domestic politics of the country in which they're operating. And in that way, I find it to be completely removed from any sort of, from any sort of kind of healthy questioning of what we're involved in.

30:37 Lars Peter Nissen

If I have one issue, I'm not sure we're totally aligned. It sounds some of the words you use or some of the things you say, sound a little bit too mechanical for me. It's almost like this was somebody's intent and therefore this happened. And I think that war and conflict is, yeah, there's a lot of intent, but it's bloody chaos what comes out of it.

30:55 Clionadh Raleigh

I think that the chaos is an accepted cost. It's what's going to end up. So, I mean, people go into conflict with, well, the people who organize conflict, go into it with an intended range of outcomes that they're seeking. Whatever, however they get there, they get there, right? So I would agree with you that the actual events themselves are both, you know, they're tragic, they're criminal, they're truly atrocious. But if you look at the underlying political logic from the beginning to the end, and the fact is that that violence was the means to move that political agenda forward, I think that things become a lot more straightforward. We often face, and this isn't a bad thing, I just think it's important to recognize, we often come to a conflict question from the perspective of the welfare of the people who have been affected by violence. How many ways in which people have been repeatedly attacked or what are the fatalities or what's the exposure rates? How many children, how many women? That's all incredibly important, especially for the work that you do. And it's very important so that we know the kind of, again, the welfare cost of a conflict, right? But that's only one way of looking at conflict. Another way is to look at it as warfare. So who's involved with which agendas, which alliances, how the network is working and how they're proceeding through a political goal. So looking at it through a perspective of the conflict itself, who's involved, how many groups, what are they up to, et cetera, that to me is incredibly informative about the political functioning of that violence. Not the welfare, but the warfare.

32:48 Lars Peter Nissen

My question is whether that sort of, I hope you don't mind me calling it mechanical sort of way of thinking. I just…

32:56 Clionadh Raleigh

I mean, I disagree with it, but I'll let, I'll let it happen. Yeah, it's mechanical. There's a business, there's a market for conflict and there's a business agenda at the end.

33:06 Lars Peter Nissen

But I think you have to be careful not to go look at what happened, what the outcome was, and then go back and link it too strongly with the intent behind it, because I think it is often, I haven't studied this like you have, but it comes across to me as so chaotic and so complex in the sense that, you know, intents are reshaped as we figure out what, it's the art of the possible war also, right? And so, don't you have to be careful that it's not about shooting the arrow and see where it lands, and then you draw the target afterwards?

33:40 Clionadh Raleigh

No. I dont, and I'll tell you what, like, let's, let's take the current Sudanese conflict, right? You had two groups, the leaders of which knew that they were well matched, right? And so they knew that in order to pursue this violence, they would have to actually do quite a lot of alliance building, especially in peripheral areas across the country. And so they started that both before the violence broke out in Khartoum, and, and they have proceeded with that throughout this conflict. So if you are either of the leaders of the SAF or the RSF, you have a clear agenda not just to fight each other, but to also seed your alliances and networks throughout the country. The violence that emanates from those local conflicts looks chaotic, but it is highly structured along the lines of we need to both deal with local front lines and of course, the front lines presented by the other allies, the other network and the enemy, you know, the other group. And they have, they're obviously thinking about like, we're best to move into how best to take over territory. But I would say that it's not chaos. It's a conflict. There's an agenda and the violence supports that agenda.

34:59 Lars Peter Nissen

I get that and I agree with what you just said. I think the point I'm trying to make probably not very eloquently is maybe it is that history is written by the victor in a sense, and that we always look at these conflicts in retrospect. And, you know, the military has a saying, no plan survives first contact with the enemy. And so you go in and then you start fighting and the fog of war, and you come out on the other side and somebody won. And then I think you write a story or you tell a narrative that somehow is retrofitted to what the outcome actually was. I think that's the danger.

35:36 Clionadh Raleigh

Yeah, I think that, so I think that the place where we're having a miscommunication is that I don't think that Hemedti said, you know, on the 9th of June, we're going to be here and we're going to engage with this enemy. I think he said, we're going to probably fight in about 12 different front lines. And so we need to build alliances with the groups that are present there in order to be able to be prepared when we come to that, right? So the agenda is I want to take over this state. The steps to that are I need to first control Darfur, I need to then control Kordofan, I need to continue fighting in Khartoum, because the political agenda is domination. It's not, you know, I just want to survive. But I would say that when we don't apply an understanding of what the function of that violence is for, then everything does look chaotic. And what we end up doing is missing quite a bit of the ways in which there is a local agenda, there's a regional agenda, and there's often a national agenda to conflict. And they either feed into each other or they end up creating alternative types of conflict.

36:47 Lars Peter Nissen

So maybe just to recap some of the things we talked about. So you're writing your PhD 20 some years ago. You can't find subnational data that gives you the granularity you want to really deconstruct the narratives we have around conflict. And so then from that emerges ACLED, and today it's an impressive machine collecting these several hundred thousand pieces of information a year. Now, tell us how you changed the world by doing this. Give us the best example you have of a narrative that you were able to tell that showed us something different about a conflict. And that that actually made some decision makers change their mind.

37:32 Clionadh Raleigh

Oh, that's a, that's a good point. I, um, I think that what we've been able to show a lot, of course, is that the number of groups that are operational in a country is far higher than what people assume are operational. And so the ways in which people need to prepare to engage in a conflict is not simply to negotiate with one group or to convince the government to engage, but rather to look at the ways in which the political system has created incentives for creating violence. So we've done that quite a bit. And I think that it certainly has fed into very senior political decisions made, especially of course, in some of our original supporting countries. I can say that there's been cases where ACLED has been used to confront countries where they claim something hasn't happened, but we have the evidence to suggest it has, and has led to changes in the battlefield as a result of being confronted with this knowledge.

38:37 Lars Peter Nissen

Can you give us a concrete example of that?

38:39 Clionadh Raleigh

There was a discussion apparently in NATO, where one of the countries that NATO was engaging with was claiming that they had no activity on a border region. And we were able to present quite a large evidence base that they in fact were actively fighting the Kurds in certain regions that they claimed not to be. But I would say in other cases, we get a lot of, let's say pushback from certain countries about how we've decided to produce the data openly and it has created uncomfortable situations for them to have to answer. Whether it's Saudi Arabia and Yemen, or it's the Filipino government and their own activity. The fact that the evidence is out there for people to use, and it's consistent and it's updated, makes people have to answer for their activities. And that is, of course, a very positive thing. One thing that I have a personal fixation on is to move the conversation around climate change and its effect on conflict past this ridiculous causal discussion that quite a lot of people, especially in our humanitarian community, tend to be involved in.

40:04 Lars Peter Nissen

It's good business.

40:06 Clionadh Raleigh

It's good business. And that's the most horrible part of it. It's not based on anything real, but there's money behind it. So people perpetuate just pure nonsense about there being a connection. What is connected is that countries and localities and communities suffer from both often, and that they should be treated as distinct and co-occurring problems. But instead, we chase our tails in workshops and in conferences talking about something that isn't true so that we never actually have to deal with these co-occurring problems.

40:36 Lars Peter Nissen

Just unpack that a bit.

40:40 Clionadh Raleigh

Well, there are ways in which the way… that how a humanitarian or how a development agency or how policies should be distributed subnationally… that to encourage climate change adaptation, which almost everybody needs to start engaging in, quite crucially, especially in low-income countries. And the ways in which that's possible is going to be predicated on the security environment, right? You know, helping to make sure that a market is functioning or giving climate information or creating a crop insurance scheme are all very interesting and useful climate adaptations, but they are going to be successful only in particular types of security environments. It doesn't mean that the conflicts that are occurring in any of these places have anything to do with the climate change. But instead of dealing with these things as, this is feasible in this environment, and this is feasible in this environment, because those are going to change depending on the conflict rates. Instead, we are fixated on this notion that if the conflict must be about the climate, and so therefore we can't solve the climate issue until we solve the conflict issue, which comes back to the political engineering that I mentioned earlier, and how worrying it is that external entities want to engage in very, very highly localized political disputes in order to be able to provide adaptation, adaptation support, which doesn't need to happen. And I think often removes the ability to actually do something very positive locally because we want to do too much in something we can't do.

42:23 Lars Peter Nissen

So what's your solution to that? Apart from producing really good evidence that shows that that's the case, what should we do? What's your advice?

42:30 Clionadh Raleigh

Well, we just wrote a piece for The Lancet where we outlined that if you had better conflict information and you knew exactly the form of violence that happened. Let's take Nigeria as a perfect example, right? The state of Nigeria has several conflicts occurring in it, and each of those conflicts allows for a certain type of adaptation to be most effective. I'm desperate to send my husband and good friend up to ISIS-controlled areas of northern Nigeria to see how they're engaging in climate adaptation. But whatever happens there is going to be very different than what happens in the north… the northwest of the country where you have control by militias versus what happens in the middle belt where you have you know you have resource competition between livelihoods that require a different type of adaptation. All of which can can receive adaptation training and funding and support but it has to be one different and two to be effective you cannot claim to want to solve the political issues in that region but rather to attend to these immediate needs.

43:42 Lars Peter Nissen

It's an interesting element to introduce into a marriage, I would say, that you ask your husband to go to an ISIS controlled area.

43:50 Clionadh Raleigh

Yeah, he just doesn't, he doesn't have like, he's not a political animal in that way. Neither is Andrew Lanky, so I thought, perfect.

43:58 Lars Peter Nissen

Yeah, I think the humanitarian experience is that your value as a hostage and potential decapitation, candidate doesn't really rely on you.

44:09 Clionadh Raleigh

That's an excellent point, right? But I would imagine they both would be extremely annoying hostages.

44:18 Lars Peter Nissen

Yeah, that might speed things up though.

44:20 Clionadh Raleigh

So what we did in this paper is we outlined what's most effective where based on some of the characteristics of the conflicts we commonly see occurring.

44:30 Lars Peter Nissen

Yeah, I guess that conflict doesn't just reshape the battlefield, but all of society and all of the possibilities to intervene.

44:35 Clionadh Raleigh

It does. I mean, even as simple as - does a conflict have significant number of roadblocks or is this a conflict in which, you know, there's a control on costs or control on the price of materials coming out? So is there agricultural control versus infrastructure control versus, you know, a closed entity like you would find around the Lake Chad region where ISWAP has quite a lot of control. All of those create different opportunities. And we have to start adapting to them if we want people to receive the support they need to deal with climate change.

45:14 Lars Peter Nissen

So I think it's fair to say that we are in a time of great changes. I think on one side, geopolitically, obviously, we are seeing tension rise in a number of different countries and regions. And secondly, technologically, I mean, artificial intelligence, I think, is going to reshape a lot of things, and one of them probably is our access to data, our ability to collect it, but also potentially AI can be a source of pollution of the information space I would say, you know, you can see how you can… somebody once described AI as having lowered the cost of producing bullshit to zero. Yeah. So what's your thinking there? Where are we going in terms of, I mean, we're in the same space…. You focus on conflict, we focus on humanitarian action. We collect data, we collect it so that people do something different or have a different narrative to operate on. How is that going to be reshaped?

46:23 Clionadh Raleigh

That's a great question.

46:25 Lars Peter Nissen

Actually, yes, and I will put it even more succinctly because I've worked with a very clever guy called Paul Skinner, and he said to me one day “you know, really it is all about how can we have narrative competency in a time of AI and almost in a post-truth situation. And I like that concept because it brings together both the need to have data and narratives that are founded in evidence and reality. And on the other hand, to have competent decision makers who are able to use this data. So how can we ensure narrative competency in the time of AI?

47:10 Clionadh Raleigh

So those are great questions. I in some ways couldn't care less about AI. And the reason is this, right? Like I'm in the information business, right? And what I know very clearly is that I get that information has now become a market and AI has lowered the cost of that as we were saying before. But the information that I provide has its highest value when it's not generated from nothing. It's not generated, in fact. It is something that we have gone out and sourced and verified and supported to be able to be done. Like more people are interested if I'm able to say, listen, we... a three-day walk from this capital, let's say sub-state region, is where we found something very interesting about the pattern of conflict within this country. Rather than we have 50,000 events of which we don't know whether any of them are accurate. And it seems to me that I don't think people need more information. I mean, I think that they need the exact amount that ACLED provides, the exact amount, but not more.

48:36 Lars Peter Nissen

But two things, two reasons that you should care about AI, I think… but it's nice to meet somebody who doesn't. One) the information space is an integral part of the battlefield. You can see that in Gaza, you can see that in Ukraine very clearly. I think it's always the case, right? But I think it's being taken to a new level. If you look at the way disinformation has been used to reshape the battlefield, for example, humanitarian actors in those two, in those two conflicts. I think it's clear that the fight for the narrative, the information space, is an integral part of the conflict.

49:19 Clionadh Raleigh

Sure, but like how much of that is based on information?

49:22 Lars Peter Nissen

Well, if it is mis- and disinformation that is being pumped out there and people don't know what it is, it is based on information. It just happens to be bullshit.

49:30 Clionadh Raleigh

I think that your propensity to believe bullshit is highly based on your preceding political positioning?

49:40 Lars Peter Nissen

I would probably agree with that, but perceptions are somehow real in a conflict. And if you can change the way people think about a conflict, you change the conflict.

49:50 Clionadh Raleigh

That's true. But can I just very quickly like, so if you're, you know, the equivalent, if you're a data set that is trolling the internet for whatever it is that's out there, then your likelihood to be picking up vast amounts of disinformation is enormous, just totally enormous. But I think that over the last 10 years, people have become very aware that if you're not, if you're not in many ways quite rigorous about the information you collect, you are collecting garbage. So why would AI change that equation? Just because it happens to be quicker and cheaper to collect garbage.

50:31 Lars Peter Nissen

Yeah. And I think because what will happen also is that, you know, it may start, people may not realize what they're in, you will be inundated in these false narratives. And so the sensible voices will drown in that sea of mis and disinformation and be amplified sometimes by people with a purpose to mess up things or further their own course. Sometimes by people who simply don't know what they're doing.

51:05 Clionadh Raleigh

Yeah. That, I mean, I think that that's happening now, of course, but I suppose like how that's supposed to influence how we operate or how we work… I mean, I don't know about you, but I am not interested in becoming an organization that, that looks for disinformation, not only just because I'll be drowned, but also because it doesn't serve the purpose of creating the good information that you and I both has spent these decades doing…

51:29 Lars Peter Nissen

I'm actually thinking that we should work more on developing products that do sort of a meta analysis of the quality of the information space overall, and not just focus on collecting data ourselves. I think helping other actors steer through a really shitty information space might be quite impactful.

51:54 Clionadh Raleigh

Potentially, I know that when we've done meta-analyses, which we do pretty often, right, about, for example, how much of our data from Burundi comes from a local source versus any sort of regional radio or newspaper or whatever the case… I would say that they're kind of useful, they're useful perspectives to have in our back pocket, but I wouldn't say that like anybody's too into them. I think that there's a lot of obvious and real concern about what AI can do. But when you sit down and you're expected to give a complex analysis to somebody, they don't want to hear the mechanics of it. They want to know that it's good, but they don't want to know, they don't want to know the methodology through which you generate this.

52:44 Lars Peter Nissen

Interesting. I probably think that, I think where my mind is, is probably around again, this narrative competency where I don't think we can just focus on the data we collect and the quality of that data. We also have to think about the other side of the equation. How do we best equip the decision makers that we serve, that are our users who, I mean, for me, our success lies in the agency of others. So somebody should do something different if we produce a piece of analysis. And I think to equip those people with better tools or to give them better literacy, I think some level of meta-analysis of how certain can we be of this, how much garbage is out there, I think that is impactful.

53:10 Clionadh Raleigh

I mean, I think that measuring garbage sounds good, but can't happen and will take up all of your time and not actually be that impactful in the end. But I think being very clear about where you got information is very useful and how you use it. So for example, in our cast, which is the prediction system that the data science team have put together, there's a whole page called, where did these estimates come from? What are the error bounds? How frequently are we incorrect or correct about our future casting for, you know, all of the countries in the world with conflict? And that seems to be the level that people want. They don't want more than that. They're coming for the conflict analysis, not for the methodology. They want to know that you've done it well and rigorously, but they don't want to be inundated with the details.

54:30 Lars Peter Nissen

So I agree with that. I'm not talking about 50 pages on how much nonsense is out there. I'm talking about a traffic light system. And I think actually AI may be able to help you in doing that. And I do think that that can be useful, especially when you're dealing with urgent decision making and sort of operational level decision making, where things move quite quickly.

54:54 Clionadh Raleigh

Oh, no, I wouldn't trust it for that as far as I could throw it. No. Especially urgent decision making. What you want in urgent decision making is somebody who knows something about the context, not whether or not AI told you, this source might be okay, might be okay. I can't confirm it, it might. I'm hesitant because I do think that there's a rush to figure out how we're going to use it. But where I can see it being useful is, for example, helping come up with or discern patterns that we might not be able to do ourselves and the consistency of those patterns that I like. But I will say that people's likelihood, not just of believing any information, but believing patterns, is closely tied with what they want to believe. You know, I've just, for example, in my constant iteration that climate is not creating conflict, there are some people who are happy to believe that the evidence shows that it's not, and there are some people who just don't want to believe it. They don't want to believe it. They have bought into this narrative so much that no amount of information is going to convince them otherwise. And I think that we overestimate the likelihood that better information or more information changes how people interpret conflict or the conflict environment.

56:21 Lars Peter Nissen

I think there are things we can do with it. And the question I have to ask myself is we spend $8 million a year in staff time on analysis, more or less. Can we get more bang for the buck? Because when I see some of the stuff people are doing, I am not sure that's a human needs to spend hours doing that. I think maybe some of these things can be done by AI. There needs to be absolutely no shift of accountability. If you produce something, you're accountable for it. You can't say, oh, the machine told me that. You have to do your due diligence, but we have to learn how to live next to these things. And then if we do that, maybe we can spend more time focusing on the right stuff. I think there's something there, but I'm with you. It's not a solution to all problems in the world and there's a lot of unrealistic bullshit going around.

57:11 Clionadh Raleigh

I, I, yeah, I agree with you. I mean, signals, you know, the new, the new work by OCHA is, is using AI, to both summarize what's happening in a context and also to demonstrate whether or not something has, you know, gone up or down. That's, that's I think a good use of it, but I mean, I wouldn't want to get rid of all your analysts, Lars. I'd keep some of them around.

57:38 Lars Peter Nissen

No, absolutely not. I want them to spend their time better. It has to be human led. No, no, no, it absolutely has. But there's some things human brains are not great at.

57:50 Clionadh Raleigh

Yeah, sure. I think there's some things that we have not as a, as a community at the moment, been, I don't think we've been rigorous enough asking ourselves certain questions that we can answer because the answers are uncomfortable. Not that the answers aren't achievable.

58:10 Lars Peter Nissen

Clionadh, it's been a real pleasure having you on the show to ask some of those uncomfortable questions and to maybe also disagree a little bit. I've enjoyed this very much and thank you for all the work you do with ACLED. I'm obviously a huge fan and I think it's truly impressive what you've done.

58:24 Clionadh Raleigh

Well, thanks very much. I really enjoy the podcast. I enjoy the disagreements. It's a real pleasure to be able to chat about them. You know?