 Hello everyone. Welcome to our 2.15 p.m. session where we will be talking about the conflict observatory program and having a discussion with our implementers who are part of that work. I am Susan Wolfenbarger. I am at the U.S. Department of State in the Bureau of Conflict and Stabilization Operations in our Office of Advanced Analytics and I have the pleasure of being the grants officer for this program. So I am joined today by some of the implementation teams from the conflict observatory program and I will let them introduce themselves in just a minute but I wanted to start off with some framing information about the program before we dive into what I think is going to be a really fantastic conversation about all the technologies and methods that the teams are using to document Russia's war crimes and other atrocities in Ukraine. So the conflict observatory program was one that we designed in CSO actually prior to the invasion last year and a group of us came together to think about what might happen in the coming days and weeks and what we might be able to do to respond to a potential and anticipated full-scale invasion of sovereign Ukrainian territory and this is really a line with CSO's mission which is really to think ahead and to devise strategies and we were really thinking about our previous work in the development of newer technologies and methods and how we could leverage those in that situation in particular. So we were really focusing on remote documentation methods in this program and eventually as it evolved the focus became that of documenting war crimes and other atrocities and so the the conflict observatory program is a cooperative agreement. Esri is the the prime on the the grant but we have some really fantastic other folks that are going to be talking to you today and we're really leveraging commercial and other publicly available data both commercial satellite imagery and other types of open source information to do this documentation. So we have representatives from Yale's Humanitarian Research Lab, the Smithsonian Cultural Rescue Initiative and Planetscape AI who are here with us today and Andrew is up on the screen. So far the team has produced about 19 public reports you can see them on conflict observatory.org and all of the work that's being done by these teams I think it's really important to note is following international standards so they're following protocols like the Berkeley Protocol for Digital Open Source Investigations so that any of the work that they are doing can be used to support international justice and accountability mechanisms. I think we've probably all seen a lot of the fantastic reporting that has come out from the teams. The the Yale team put out a report in February on the relocation of Ukraine's children into a variety of locations across Russia and we've seen work from the the Smithsonian team on looting and destruction of cultural heritage sites and the Planetscape AI team is doing some really phenomenal work with AI and machine learning to help us really cope with a scope and scale of documentation that we haven't had to think about before. So I'll just wrap up by saying like we see this as a it's a capability that we are using in certain ways in Ukraine for war crimes documentation but this is you know as I said a capability it could be used in other types of situations and we're really hoping to be able to apply this innovation in other places in the future. So with that I am going to move us into our main discussions so each of the leads from the different groups are going to give you about five minutes to talk about the work that they are specifically doing as part of the conflict observatory and then we will move into some questions that that I've already prepared and definitely I would love for everyone in the audience to think about questions that you might have for the implementers as well. So we're really hoping to have a conversation about you know the the work that they're doing the the technical documentation but a lot about the difficulties and things like the trauma that might be associated with researchers and all of these things so we're really excited about this conversation today. So I will just kick us off and by virtue of who is sitting closest to me we'll start with Dr. Katherine Hansen. Thank you Susan. This is on right? Okay. Delight to be able to join you all thank you for the introduction and it's really a pleasure to be able to tell you a bit more about what we do for the conflict observatory. Smithsonian is just one part of the Smithsonian team is comprised of several different organizations that work work on documenting cultural heritage damage and so probably what I should preface this by saying is that we're really a research team we're comprised of a group of scholars who are focused on doing systematic documentation with the purpose of better understanding how and why cultural heritage is being damaged in the conflict and I'm going to pause for a moment because often I need to explain why cultural heritage might matter in a conflict why this is something that should be documented and I just want to clarify that by saying that it is it is one part of the fabric of society and cultural heritage is is often particularly for the built and physical cultural heritage it is component that really represents a people its diversity its history and its identity and so to that end we are not obviously the first ones on the ground by any stretch but cultural heritage practitioners are certainly there in crisis and conflict situations documenting what's happened. There have been war crimes prosecutions for damage to cultural heritage in the past and one of our goals in this situation is to make sure we're documenting everything to an international standard where it could be used as evidence. So Smithsonian has a group of researchers who interdisciplinary team primarily comprised of anthropologists, archaeologists, art historians, and political scientists and we have a group from the Virginia Natural History Museum's cultural heritage monitoring lab who work with us and then we also have a group from the University of Maryland's international conflict studies program and those the academic sides from both of our major fields put together are really bringing the open source data sets that we have both the conflict data sets and then the open source information we're able to get and then our own data set that we're creating of cultural heritage damage events paired with the remote sensing and so we've had some really collaborative work that we've done with Andrew's team and we're also bringing in a lot of work with high-resolution satellite imagery and infrared remote sensing. I think at this point I'll turn it over to Caitlin and Yale's team. Thank you so much Catherine. I just want to reiterate how critical it is that the work that Catherine and her team is doing though work that we highlighted most recently in our report on the forcible relocation of children to Russia that really one of the aspects in that report it kind of gets a little overwhelmed by the fact that you know children are being taken from their homeland to a new place and in many ways that they are in some areas they're being put into foster home situations some are being disappeared through adoptions their identities are in fact being erased through renaming through their original identity numbers their Ukrainian identity numbers being replaced with Russian identity numbers and other measures to essentially try to strip apart the ways in which we might be able to track them from one place to another make it harder to find them in the future make it harder for their parents and their families to find them in the future that in itself is horrifying enough the really wide scale and systematic event though that we tracked and that's what we do at Yale Humanitarian Research Lab we document wide scale and systematic atrocities in war and the way that we did that was we were looking at the level of what's called reeducation it's really a manner of indoctrination and that is at its heart it's about not just erasing but replacing culture you can't do that if you don't understand what it is and it's actually very hard in many ways to appreciate if you don't have the anthropologist the political scientists working side by side because they appreciate the depth and the violence of what is taking place better in many respects than others do so at Yale a lot of our work is done in sort of two modalities our most forward leading modality is to detect collect analyze and document and securely archive all of this evidence and we look for all sources of evidence anything that's open source by which we basically mean anything we can find that is publicly available it's unclassified it's there for anybody to find it it's often digital it's out on the internet in some form doesn't mean it was originally digital but it's been somehow captured in a digital format so somebody may have taken a photograph of a road sign for example or there's a piece of paper that somebody is scanned we try to capture all of that process it make sure that we then fuse all of that data together in various formats we also do this with some fairly complex and sort of high-end databases and that includes very high resolution and to like high medium resolution satellite imagery to make sure that we can place all of that information in time and space which is crucial data is nothing if it doesn't have context and in doing all of this our first goal is to make sure that we can support the most like near real-time the most forward-leaning decisions that need to be made by policymakers by civilians and in that case by parents who are trying to figure out what to do next that's the most urgent task the next most urgent task is to hold the perpetrators accountable for their actions and that's why as Susan said we do all of this to these standards and it's so important that we know what those standards are in advance so that we can design all of those things with those standards in mind if we don't do that part then we may inadvertently destroy key parts of data that can become evidence in courts and in other accountability mechanisms later on so we have to be very careful about how we do all of this how we capture it how we securely archive it and we are constantly going back over our records back over our collections in our archives to make sure that we've done it correctly so that's a lot of what we do at Yale I also just wanted to note before I hand it to Andrew that we're excited to keep this conversation going all the time I did forget to tweet before we started but you'll find me both individually and at our Yale HRL handle so we'll follow with questions for everybody who's following on YouTube I hope there's more than just my mom online but I'm on all right Andrew over to you hi um are you guys able to hear me all right great um well first of all I want to apologize for not being there um we're a small group out on the west coast and because I'm heavily involved in our daily operations even getting away for 48 hours isn't quite feasible with all that's happening um yeah my name is Andrew Marks I'm the CEO and chief scientist of Planscape AI prior to founding the company I was a professor at the University of Southern California specializing in spatial computing and my company is pretty much just me and a small group of my best graduate students and what we're doing is leveraging advancements and artificial intelligence and high cadence data to inform humanitarian and human rights work worldwide specifically for the conflict observatory along with the illness masonium where we collaborate on authoring reports but uniquely we also provide a couple of specialized analytic services so the first service it's what we call the destroyed building analytic service this employs advanced AI and daily commercial satellite imagery from planet labs you know all commercial data to automatically detect destroyed buildings across Ukraine so every day all buildings in our study area are evaluated by our AI if they're intact or destroyed in quotas as such this is providing an unparalleled forensic data set of when and where buildings are destroyed throughout the conflict this analytic service has been crucial in various projects including tracking the destruction of marioble and corroborating eyewitness reports of the bombing of civilian infrastructure and the second analytic service is the human mobility metric and this provides a macro level understanding of population movements across Ukraine on a daily basis the service leverages data from multiple sources to provide a comprehensive understanding of how people move this has been especially useful in the eastern regions of Ukraine where reliable data on migration is scarce this service has been utilized in several conflict observatory reports including one with the universe of Maryland which demonstrate that Russia's armed forces are the primary driver of conflict related displacement not just the conflict in general so in order to leave time for plenty of questions I'll close with that and saying that we're very proud of being part of this effort and I would congratulate Susan for the excellent work she's done in assembling the conflict observatory great thank you thank you all for your opening remarks I have a lot of questions for you that I hope will help spur some thoughts amongst the audience but maybe we'll start with you know we're talking about your current work but a lot of this and the reason why you're part of this program is because your past work in the conflict space you've been doing a lot of past projects on remote monitoring on open source documentation and evidence creation so I thought maybe a first question would be if you could kind of talk about some of the lessons that you learned in your past work and how that really helped in the development of this project what has changed I know I've been working with you all on and off for more years than we might want to admit right now and so you know some things have changed so I think you know hearing about you know how the field has moved is a great starting place for us and I'm not sure who would like to to kick us off sure so the first major effort that I worked on was Sudan circa 2011 this was focusing on how primarily the Sudanese armed forces which you might be hearing about a lot of it in the news again this is like the main military for Sudan and its government was systematically targeting the mainly the Nuba people but a variety of different civilian populations of these three contested border states along the border with South Sudan and just to be clear these were Sudan's own people these are not South Sudanese the thing that was different then was we were working in a very similar modality sort of near real time trying to get material out as quickly as we possibly could for to give you an idea of how quickly we would put a report out sometimes we would get satellite imagery in by like 11 a.m. and have a report out by 1 a.m. the next morning okay so that quick of a turnaround timeline and they would hit they were they mattered they were they were getting picked up they were making a difference but we barely had guidelines at all for how to do what we were doing and it wasn't because we didn't look for them we looked to journalistic best practices for how to do what we were doing with the most ethical possible guidelines we looked to you know data practices and in academia we went to our you know this was work that was based out of Harvard so we went to Harvard and we were like okay do we need a institutional review board to sort of keep us in check or to like you know touch base with us from time to time and they were like you're using open source material and satellite imagery you're not collecting any personally identifiable information therefore nobody can technically be specifically like harmed by what you're doing and we were like oh they can definitely be significantly harmed by what we're doing that's not true at all and they're like yeah but you can't like make a direct liability claims like you're good and we were like oh this now I know what data colonialism looks like and it looks like me um so now all these years later we we spent the downtime we had after the initial 18 months working on that kind of going okay now that we you know now that we've done that um in a very scary like you know speeding down the highway at 150 miles an hour with your lights off and it's pitch black outside and you're just hoping you're on the right side of the road mistakes were made those mistakes cost lives I personally hold myself responsible for at least two lives lost and others very severely affected and that's not because it's not just because I'm Catholic but it's it but it's because I know exactly what I did and I know how that information was used to target those civilians and I'm not trying to beat myself up with it like it's just the reality so we do better now because we then spent the better part of the next eight years learning from it intentionally learning from it raising money so we could dedicate time to go talk to everybody we could talk to you about what we could do better and then writing about it and documenting all of it including documenting the loss of life documenting all the things we should have done differently and then figuring out how to choose better partners so it wouldn't happen again that's how you lean into the suck and you do it better and we'll make different mistakes this time but we're not going to make the same ones I think very much the same in terms of when a lot of this work began there were not a lot of boundaries established guidelines no established guidelines and ethical conversations that were really mostly among ourselves and I'm going to just go ahead and date myself in 2004 when I was working on beginning my work on my dissertation is when I first started using satellite imagery to document damage to cultural heritage sites and we had it was the commercially available stuff it was over southern Iraq it was horrible refresh rates right how often can you actually get that great resolution and and a real challenge of of me looking at this being like well I guess I just count these holes now and I'm just going to have a bunch of other research assistants count count things and the dramatic difference now and having automation of any sort of having so so there's there's the the the ethical frameworks and the guidelines and that's been a very big change and then there's been a huge technological change and what's available and so when Susan first talked about what conflict observatory would be I remember you you sold me on the line that well this is what I'm building what we would have wanted back in 2014 and I was thinking to myself wow well we wanted everything back in 2014 and we didn't have very much so this is this has been so cool to see when when we were working on damage to cultural sites in Syria very little was automated if anything there were very few guidelines there were there was also this pressure to attempt to compete with news media and that's another point that I just want to kind of raise and I think that we have moved beyond as research groups and I think is it's something that we still struggle with a little bit when we talk to funders when we talk to when talk to media but that we are not in the business of breaking the news that's that's journalists jobs and they do that very well our job is this research and our job is to systematically document the damage in a way that it can be usable and I think that having that clarity has also that that's been really crystallized since you know we we started looking at this stuff early on and and has been especially clarified since 2014 I'm conscious of the time and I want to make sure I give Andrew a chance to weigh in on this one great thanks I think like the rest of you I did in my dissertation using remote sensing looking at Darfur and a lot has changed back then I was using a Landsat imaging every 16 days at 30 meters now we use planet images which is every day at 3 meters so that's the first of three things I think the really change the second is the computing instead of me running for MacBooks 24 7 now we have distributed cloud-based processing that pushes through tens of thousands of images across all of Ukraine overnight so it's it's amazing but the biggest thing that has changed I think is the audience it's everyone who's there to meet this conference it's people in government and in international organizations because folks really realize what is possible and are increasingly receptive to new data new analysis approaches and new ways to document atrocities and other issues thank you great thank you all and I really think we can't understate you know how much things have changed across the board since a lot of us got our start on in this field so with all of the the technical advances will start there that have happened you know Andrew was just talking about you know going from 16-day Landsat 30 meter to daily 3 meter uh planet scope we have all of this information at our fingertips now so how are you deciding where you're gonna look where you're gonna research you know to what extent are the things that we're concerned about in uh Ukraine documentable with our methods and so I thought that might be a good next piece of conversation to think about um maybe Andrew do you want to do want to start with this one first of all I think and this this was really highlighted in a collaborative report we're working on with Catherine right now is that you've a similar to great team Susan I'm not a cultural heritage expert and so for example there's an upcoming report about Donusk and we were able to see that it looked like there was some kind of maybe disproportionate damage to cultural sites but you know some folks at the University of Maryland were able to apply some much more rigorous methods I was looking at I think it's a weighted two-way fixed effect quasi-binomial model with time trends and then the Smithsonian all their expertise really weighs in on what this means with the cultural buildings so um I think for me I'm just conscious about staying in my lane doing what we do well which is monitoring large areas and providing forensic data sets and then otherwise we work we make sure we collaborate on any reports that come out and yeah we we definitely appreciate that creation of data that you're really serving as the role on on the team so that you know the the other teams have a whole another new set of data that doesn't exist elsewhere to base some of their additional work on so maybe you all can talk some more about that honestly Andrew I I think we haven't fully taken advantage of the data set that you all are creating and we're only beginning to scratch the surface surface of what we can do and a lot of that is because to get to Susan's point the bulk of what Smithsonian Smithsonian Cultural Rescue Initiative does for conflict observatory or elsewhere is at the request of local partners so we are investigating researching providing information at the on on topics that have been requested specifically by our colleagues in Ukraine so the Ukrainian Ministry of Culture and then also Harry which is a heritage emergency rescue initiative there as the NGO and those are the two big consumers of what we produce and so the very first product and product research paper series of research papers that we make um our potential damage reports so they aren't confirmed they're things we use based off of NASA's firms open source you know fire detection infrared remote sensing and Rodin our script team member did that pulls that daily so we do a monthly report on an overlay of any cultural site that looks like it's within a one kilometer radius of one of those impacts and then we have to remove agricultural burning areas and then often um often those become the most important research areas for us because that's our list of potential damage that's what we then share with our colleagues who can go out on the ground and check um if we're getting a lot of stuff showing up saying these remote areas which you don't have communication with anymore look like they're being damaged we want to make sure you have that information at your fingertips um and sometimes that's working and sometimes that's that's not uh as as useful as it could be um and similarly we uh we have an upcoming report on looting uh coming out of 12 looting reported looting events that took place and we're with this is primarily an open source scrape um of information from our our data set paired with uh Russian territorial control data sets and when museums and cultural institutions were looted and when they were not in their location of looting and um in that that request came in both from our colleagues in Ukraine as well as also you know there's been a lot of interest from other agencies specifically law enforcement agencies as well um so we we primarily uh take requests from our colleagues in Ukraine but obviously there's there's other interested parties who who receive the information as well Caitlin was there anything you wanted to um I would just say we also uh work based obviously based off of quest as our colleagues do the the other key thing that we factor in though is that we tend to look at our in terms of narrowing our scope um I'm sorry because there isn't really a way to say this that doesn't sound super creepy so content warning um we narrow it the way a perpetrator narrowed was it we look at civilians we look at the we we look at what our mandate is and then we focus in on what are the targets of opportunity the way a perpetrator does and we look at what's rich what's available um and what is primed for attack um and we look at to see what the indicators are that that sort of says you know this is what's being lined up the most important thing that we keep in mind as researchers is that atrocities happen in slow motion they are very very easy to see coming once you kind of see all the logistics that have to be put into place especially for mass atrocities um they don't just happen overnight that I think you you especially see with our last report where when you look at just if you just think about the logistics alone the number of locations that are involved how big they have to be to house that many children um the transit logistics the identification of the kids the recruitment involved how many people had to be involved in that recruitment how did those people get identified how were they all communicated with financing a lot of ever goes into all that so when did all of that start in order for all of that to be in place in the way that it did so very quickly as the war was kicking off that's the kind of thing that then shows you that you actually have to keep going back and back and back before you find the first little strings that you start pulling and unraveling for uh with you Caitlin and Catherine and your response just now you mentioned um some partnerships and I think it would be interesting to talk about from each of your all's perspectives and Andrews as well the both the coordination within the team was are there interesting things that are coming out that you aren't expecting and then also a little bit more about you know as time goes on the partnerships that you're forming uh with folks that are on the ground in Ukraine uh sure so um we we're talking this morning in a briefing for you guys about how we do social science research and um you know I I've done social science research under that umbrella for for some time this is the first time we've really done it with an interdisciplinary team and universities when I was was in academia this was something that people give a lot of lip service to being great to do but to actually see it happen on a research team where I have colleagues who are scholars that specialize in very detailed charts that I barely understand um and then colleagues that specialize in the art historical content of well what would they be taking out of an art museum and exactly how many what would you be putting in that many trucks that you can count in that satellite image and so to catch that range um it's been a really useful dialogue uh something that I'm very grateful to have the opportunity to hear what they each then contribute uh and then sort of criticize about what we're what we're touching on with with each report in our most recent report on looting we don't have a single new satellite image uh which I think I never would have as an archaeologist I never would have ever written a report about looting of any sort without a satellite image uh to back it up but because we're able to take the open source reporting and the data sets that our polycyte colleagues are able to analyze we're able to to um say something for areas that we don't have imagery for which has been really really useful um I think one of the big challenges uh that goes with that though is making sure that we're actually speaking to each other and that we're able to um be able to share our data and information more broadly uh in a way that isn't just academic and in a way that uh in this digestible and in a way that uh our current report for instance our looting report that we have coming up um it's a little bit of a Frankenstein of a report because we have two very different methodologies smushed together and I am hopeful that future reports we're going to kind of make this a little more uh enmeshed a little more seamless that's that's my hope I don't know okay Lynn how you guys feel about so I'll say that what there's um there's some great opportunities that we've had uh especially working with Andrew and his team on especially with like widespread and systematic attacks on medical facilities that's been an incredibly important one um there are some others uh there's some other specific research uh opportunities that we're thinking about on our team um but I'm not going to talk about them because they're good and I'm not letting that cat out of the bag before I have to um I don't give away stuff to Moscow unless I really need to um I will say that for uh one of the the critical sets of partnerships that we're most excited about are those with uh Ukrainian partners and frankly those are some of the folks who are doing the most essential work um because they have access in the field that we by design do not have and this is something that often comes across as a criticism of our work but I really what like it's it's important that we do everything that we do remotely and also it's really important that we do whatever we can in collaboration with local partners who can do local verification you know that verification and field follow-up um we are intentionally not trying to get into their space and we very much welcome those kinds of partnerships wherever we can access them so if anyone's hearing this and wants to work with us let us know um but we have we have a few that are coming to starting to come into fruition that we're very excited about Andrew did you have any insights on the partnership that you'd like to share um I would just say that within the conflict observatory Smithsonian Yale Esri my small company we all approach our work from different angles and we all have very different organizational cultures but I think the thing that you unite us all is we're very passionate about this work and we find ways to make this happen and I think you know I think we had we had to start from that position that this was going to work and then we've been able to kind of figure out how to do it so we're kind of pivoting a little bit away from the the research itself and thinking about the more the idea of what you all are analyzing and documenting and looking at on a on a daily basis is a very difficult and heavy topic and I'm asking this partially for the benefit of the audience because I know somewhat the answer to this one but you know I think my question is how are you starting to deal with kind of the you know the trauma that you that you and your research teams might have as a part of doing this type of work I wish I had an easy answer for this you want to go first yeah there are no easy answers um that's the answer um there's also you know for I will say the work that's done by places like the dart center for journalism and trauma work like they're excellent um Columbia does great work on this that being said um I wish there more were more public listings of great actual therapists that were readily available in everyone's region cross-referenced by insurance like that like that's the practical resource I need for my team and I don't have it you know going through psychologist I'll be just real talk I've got a team that's all that's you know we're consultants I go through personally I let go through psychology today trying to find people who can like actually set an appointment with my team who have the skills and the experience necessary for what they are dealing with with and I can't like I struggle to find that for myself let alone for all the people with my team and we all a lot of us work remotely so that's a lot of different places to go looking and there is no one-stop anything for people like us period the best thing I can basically find that comes even close to that are services for veterans and I think if any of you know anything about services for veterans you know that they were very overtaxed as it is so that's not really a great answer um we have we have to do better at this the one thing you know we at least thought about it when we thought about building our budget and figuring out how to try and provision some resources for everyone we have regular sessions as a team to talk about everyone's mental health to check in with each other and to try at least to make sure that it is understood that your mental well-being is valuable and precious and needs to be protected and that if that includes you being given you know a lifeboat to go somewhere else we will build you that lifeboat that everybody has a future outside of this work will help you find it and make sure that it is a good one for you so that also has to be part of it it can't just be like a crash out you know because that too is incredibly traumatic and I've you know I've experienced that myself but yeah there's no easy answer to that question but one thing that actually I'm going to send you an email about later uh we need to get back in touch with our colleagues from Berkeley uh because I do think that they have some materials that they've worked on that that address this um a little bit uh that that could serve as a resource because there is as Katelyn said very very little out there um and particularly for those of us who have junior staff junior researchers uh coming up you want to make sure that you're giving them uh the support like Katelyn mentioned um but also the resources and the tools um because it's not uh unfortunately it's not like any of this is going to go away uh so it's something that finding finding a way through it um is pretty key I think the other aspect which you didn't ask about but I'm going to take the opportunity to uh say just a word on in terms of safety is that um we have gone as scholars quite uh around quite a bit about uh do we want these as scholarly publications where everyone's named and for our Ukrainian colleagues who have asked to join us in these publications does the value in the weight of having a scholarly publication outweigh the potential risks that come with it and that has become a really interesting conversation that frankly uh our approach as the as my approach as the team lead and and and everyone has been pretty much in agreement with this um has been that it's up to the individual to decide um so if you as a scholar want to be named on the publication we name you um and uh I think that that has been one of the more sobering conversations we've had certainly has been will of course I want this academic credit of course of course I want this because this helps me in my future scholarship my future job possibilities and so when we think about collaboration um particularly for our colleagues uh in Ukraine making sure that if we're we're citing them if we're uh using if we're we're asking them to contribute research that we're able to give them academic credit um and that we can have have an accurate conversation also about the risks that are involved yeah but isn't that also an opportunity for us as a sector to try and evolve and say you know papers can live indefinitely time periods that like time can elapse when people can get to safety why not have an option for their name to go on it at a later date or when right when they have when they have that option that's really what we're looking at for our team which as you might see like there's only Nathaniel and I that are public um for the for the CEO at Yale yeah and yeah we face the same question but I would love I would love for you to know all the people who are the lead authors on these reports it's not us so I think we disagree on that because we've had a lot of recent experience with both Afghanistan Iraq and Syria um where we had a lot of problems with J-1 visas and the way you get a J-1 visa the way you get somebody out is for them to have had academic publications so I think when the rubber meets the road um for right now the way the system works unfortunately we have to we have to have publication credit out there um so but we our teams do it the approach differently but um I agree that it's it's an area that really has to have some sensitivity and Susan um I would just say just say from our end um you know while much of the work that we do is computational leveraging AI you know I've a group of folks that have just finished grad school spending all the days looking at cities that have emptied out or craters where buildings used to be um so while we're very mindful of mental health um for us it's really critical that we keep connected to what we're doing which is trying to reduce human suffering and document atrocities so it is a balancing act you know we we have to remember why we're doing stuff when we start each of our weekly meetings with struggling with the conflict and understand what's happening and the role we're playing to reduce human suffering so I'm going to ask one more question before I turn it over to the audience so please think about any questions that that you might have uh before we uh turn to you so I think just kind of a wrap up question for for my perspective what advice would you give to other either individuals or teams that are trying to get engaged in this uh type of documentation war crimes atrocities human rights violations what what kind of advice might you want to give them based on you know the situation that you guys have all been kind of living in for the past year I'd say regardless of where you are any anywhere in the world any conflict any stage of development for you and your team set your standards be open to evolving them but set your standards and then stick to them the worst thing that I usually see are people who set a standard and then immediately backtrack on them due to usually it's time pressure and they'll buckle under that sense of well we have to get it out right now because something is bearing down I'll tell you what like that is always going to be true there's always going to be more time pressure and it will always be in service of you compromising on on your fundamental principles and usually on principles that are keeping your findings safe and fundamentally your team safe in establishing with accuracy that what you are putting out into the world is meaningful and accurate to the best degree that you can establish it by because it meets your standard so once you compromise on that you're then you're kind of in the weeds so yeah whatever protocol you're going by you got to you got to really just make sure in advance like set a back plan to it and be like okay can we can we actually get there I think that's you know and for what it's worth that's that's really hard our own protocols are really hard for us to hit we're constantly working and coming back to the drawing board and figuring out how to do better so being open about that that's that's fine that's just part of the process but you've but you do need to try and set that to attain to it I would add that for the advice I would have wanted myself to get to is do the do the outreach to both your local colleagues who who are there on the ground and also the international organizations even if you have no intention of having long early morning conversations with those international partners over and over again that that knowing the lay of the land having an understanding of who else is working in that area who then you can contribute towards because even if and this is I think the thing I really wish that we had eventually we came to it early on like 2014-15 but it took a while there can be duplication of effort funders hate this but lawyers really like it so there can certainly be duplication of effort from a research perspective your methodology is always going to be a little different than somebody else's and if what you're doing is replicable you should be able to have someone else doing something similar getting a not too dissimilar result and so there's a couple of international organizations that are quite concerned about duplication of effort for cultural heritage documentation and others and really as far as I'm concerned and I wish I could have really told myself this back in the day and to anybody else who's starting out it's okay for there to be duplication of effort it really is this is in the end evidence and it all goes towards helping one one point Andrew do you have any last words about any advice for folks sure I think I'll echo what Caitlin said which is whatever you do you need to maintain the integrity of what you're doing the human rights community especially in academia is very small I think we all review each other's papers regularly and we've all worked together for a long time so I would say just you know be transparent clear and whatever you do do so with integrity because it's a small community thank you all so much and thanks for humoring me with that peppering of questions on all kinds of topics I want to open it up to others that are in the audience and see what things we might have missed in this discussion so far and it looks like we have someone right here I'm Ursula can it's a lot of I'm with the Friends Committee on National Legislation the Quaker Lobby and my question is we were so excited to see the conflict observatory get stood up so quickly with so many fantastic partners gathering all this important data and the conflict in Sudan just reignited and there's crickets so how do we get this kind of energy both from the USG to fund it and connect people and get it moving on all the other crises that aren't in Europe and how do we get this a low enough cost on one hand to make it easily replicable over many areas but just the political will as well to spend our money on it well I was hoping I wouldn't get any questions you know I in my opening remarks you know I did talk about how you know this is a capability you know that is amenable to other conflict situations you know I think from the perspective of our bureau this is something that we would like to be able to to deploy in other locations and you know I think we've been able to really show in Ukraine and taking advantage of kind of the the the situation and the willingness of folks to focus on it to pioneer and pilot a lot of things that we've never had that space to do in the past and so you know I think that that has been really important for us to get that you know you create that first instance of the program and then you know when you have fantastic implementers they're doing this type of research that are proving you know that this type of work can happen completely in the open completely unclassified then that gives us a space to think about okay well where where else should we be doing this where else can we leverage these things and you know and and from that point I think you can kind of you know take advantage of the resources and the investments that you've already made and you know the the next program doesn't cost you as much because you've you know you've already stood up these these platforms are are not cheap to build and so I think you know as you're thinking about additional iterations of of something like the conflict observatory it becomes easier to to do that next that next program and so you know for for my perspective as a as a former analyst I and you know the manager of the current program I really hope that we can find additional ways that we can leverage what we've been able to create in Ukraine into other other situations. Hi I'm Kate with Rosam for Ukraine. I was actually curious to hear a little bit about if you've partnered with any organizations on the actual accountability side of things in terms of using the data that you're collecting to actually work on accountability processes international tribunals working on sanctions advocacy anything like that thank you. Yes can I say more okay yes we are so our team we are in touch with various like various offices in Ukraine so we're well in touch with the Office of the Prosecutor General that is a I'm going to say near daily contact for us at this point and we're so we're responding to a number of various inquiries helping to develop a number of cases that are in motion as you know they have quite a few so that's we are far from the only ones who I'm sure are engaged in that but yes so there so at the local level in Ukraine yes and I think I'll yes oh and also so with OSCE with the Moscow Mechanisms we submitted so we responded to Moscow Mechanisms one and two the first one was actually our kickoff report and we are probably going to be responding I believe there's been like some murmurings about the next ones to come out but that we will almost certainly be responding to once they're announced for those who are not necessarily familiar Moscow Mechanisms are issued by the country so that's Ukraine sort of invoking and then we joined with the committee of experts to then fully respond and sort of be part of the overarching report and two trolls were a little bit you were directly cited in the first Moscow Mechanism report with your documentation on attacks on healthcare facilities so that was a really fantastic lead-off report from the Yale team yeah we're not mad about it hello I'm sorry hello I'm Saba I'm from Pakistan and I'm a peace linguistic scholar my research is in peace linguistics so my question from you guys is has your organization ever thought about you know working on this peace linguistic that is let me tell you about that little bit it is all about like using language peacefully by the diplomats particularly because it's always the language that you know any conflict starts with and it's always the language in terms of negotiations and table talk that the conflict ends up so like currently I'm doing my research on the diplomatic communication that has been used for you know by different countries regarding Taliban regime in Afghanistan and while I was collecting the data and doing the analysis I realized that actually the problem is with the language so I just wanted to ask about or maybe I wanted to add it add it to your conversation that what do you think like your organization has done anything before in this regard or is planning to do anything thank you that is a really important comment and thank you thank you for that I I don't think that we have ever looked at that specifically I can say that the anthropologists on my team would be particularly interested in talking about that more yeah it's as a perspective it's one that's really important I will say we're very mindful of our language and our reports and our framing it's something that we are always very careful about as we think about how our language is going to be utilized by others we probably take a very um a very sort of perpetrator target specific approach in our reviews so one of our like pre publication reviews is what we call a red team where we produce an alternate analysis basically just imagine that your report is going to be taken apart piece by piece by the worst person possible somebody who is going to fully weaponize it and then looking at it in that frame try to imagine okay how can I edit this in ways that will help to reduce its weaponization potential that can be very that can be very difficult sometimes it's a matter of relatively simple reductions or other sort of harm reduction techniques for overall protection you might notice that like there are certain sources that we redact we do make those available for journalists and for other researchers who are validated so that we want everybody who needs the data and has a you know a reliable like need to gain at full access but but it is challenging I think that may be looking with your perspective in terms of how is this language also advancing peace process would be a very interesting one to use my name is maven and I'm with the halo chest we're the world's largest humanitarian demining organization and we've been active in ukraine since 2014 we are we have teams on the ground actively demining right now but we're still looking at potential years of harm to civilians from landmines even after the conflict ends so kind of my question to you is do you anticipate your data collection on these potential harm civilians by these atrocities continuing after like an official cease to hostilities are you tracking like long-term impacts realm so we had a project for northern iraq and a long-term commitment to working on some cultural sites requested by the ministry of culture there and demining was our biggest struggle after the cease of hostilities it really it's a really really big issue that I don't know exactly where we're going to be when when that takes place but I can tell you that based on our past experience it's certainly something that we're including as we consider what happens with with sites on the ground Andrew is this one that you're planning on taking forward okay let me direct that to me yes yeah I think if we're able to if we're able to we'll continue our monitoring large-scale of ukraine for anything we can like that you know it depends on our mandate and the co but you know that's really something we're able to do is that kind of large-scale monitoring I'd say for our part our goal is to make sure that our data is as interoperable as possible with efforts like yours so that those who are carrying things forward in a post-war context and redevelopment are basically able that this is one of the things we're working on right now with actors like mnemonic and others to make sure that we're we're designing the literal databases themselves to be as functional so that no one's having to redo the work redo the methodology as Kate said every time we all do things a little bit different but if our classifications of our taxonomies are like enough that we don't have to redo that piece we're hoping that we can transfer our overall OSINT so that our open source materials is sufficient that we can say okay here's everything that we have on mining activity here's as granular as we've been able to get it and you would be able to ingest that in as swift a manner as as possible so that even if we didn't have the mandate to continue other actors would be able to take it forward. Hi my name is Maddie I wish I had something more fun and cheerful to ask but I wanted to ask about sexual violence in Ukraine I've done a little work with Dartmouth's political violence field lab and they've been kind of doing little open source work to track sexual violence and find patterns and kind of try to determine causality which has been difficult obviously as you know and so a lot of what you talked about is kind of tracking and archiving these this data for future war crimes trials to try to hold these parties accountable and I'm curious bearing in mind the kind of ethical considerations of this data collection especially with sexual violence knowing that this is kind of the most sensitive data you could find do you think that there's ever a possibility of proving that sexual violence could be used systematically or kind of an organized weapon of war is the burden of proof kind of too high in what you found if that makes any sense like would it ever be possible to prove that sexual violence is systemic or could it be used in like a ethnic cleansing or genocide trial in the future sorry that's kind of convoluted I can clarify if needed. Thanks I'll take that one um no I don't think it's impossible there are a lot of organizations that are doing this documentation better than we are right now so we're not the best folks to speak to that particular aspect that being said in principle no I don't I don't believe that it is impossible and I do believe that there are components of that that would that do lend themselves to demonstrating widespread and systematic it's been done before and it can be done in Ukraine the challenges are I mean as you know the challenges are so many there's a lot of different perspectives that we can kind of take as investigators looking at units looking at specific sort of conditioning that may be taking place among different actors among the perpetrators looking at communication patterns there yeah I think that trying to be as creative as possible with the investigative techniques and data collection that's going to be I mean that's going to be really crucial I've talked a lot with Lauren Wolfe who did a lot of this work previously in Syria in other contexts with UN women and so the yeah but this is this is the kind of thing where it's always important to keep in mind whether it's gender-based violence and sexual assault a variant on that when it comes to targeted violence against queer communities there are so many things that often don't get sort of highlighted for their potential to be documented by teams like ours but absolutely can be it's just not considered view because it's not considered viewable via satellite we don't always think of it first and then it turns out to be something that you absolutely can map anything can be located in place in time. Ziyad Ashkar from the Carter School we started this conversation today about how the field has evolved so far and wanted to ask what is your perspective on where the next step is what are the new potential capabilities where you hope the field is going to evolve to as a community of practice and research. That's a great question and that was actually going to be my closing question so good job on that and we've got about five minutes left so I definitely did not plan that with you but I appreciate that. Andrew do you want to kick us off? Sure one important thing that we're going to be able to start doing by scaling is not responding to conflicts but preemptively monitoring areas at risk of these kinds of crises. For early detection early detection means we may be able as an international community we may be able to mitigate some of what happens we might turn you know we want to stop the fire before it really gets started and so I think that's for us that's a direction that we're looking at is how do we scale what we're doing in Ukraine to a worldwide human observatory. How do we do early detection of things that might be you know that might indicate something is happening and then we can leverage other assets to really understand what's happening there. To build on that I think for us we're right on the cusp of being able to do a little bit I hate to use the word but predictive modeling to help colleagues best understand what it is they should be putting efforts in towards protecting you only have X number of sandbags what are you going to protect the monument or the church and we're just at the edge of starting to be able to do that and I think that as we keep going hopefully in the next year I'll be able to say that we're being able to to give some answers about where you should put resources to protect obviously this is conflict dependent regional dependent that sort of stuff but I think that's that's one big big thing that we're we're looking towards hopefully soon and then more broadly cultural heritage in general is a relative newcomer in this space we the big cultural heritage organizations still often think about satellites and the data associated with them as here's a big sexy picture from the sky and what can what can we do with this and well that has its time and place it's it's just the beginning right and and so trying to get sort of the larger cultural heritage community the museum communities understanding what it is that can be done with a sort of data I think is going to be really hopefully the next big next big thing that happens I think this is something that we feel really challenged by all the time right you know how where is our next big evolution where does it need to go I tend to be a little less on the AI side except in one area and even there it's very I'm more of like Tim Nick Gebru on this one and Meredith Whitaker where I want to see it go low and slow and very very intentional and that's mainly with language because everything we need to do needs to be about preservation of language and preservation of culture and the destruction of those length of languages is a very clear and intentional part of a lot of the things that we see happening it's often actually one of the very earliest indicators that it is going on so as we need as we're doing our work better I think one of the things that we also have to be mindful of is that we're not it's also this is also one of the things that we probably do the least well at is making sure that our findings our reports the as our colleague pointed out the way we actually are talking about what we're doing and what we're finding um that we are putting it out in as many ways in as many languages as it needs to be in so that we can be having the conversation as fully as possible in real time and not constantly having to ask other people to come to us any comms professional will tell you that's not how you have a good conversation you always go to your audience so how can we ask a lot of members of the global south right now who are feeling very alienated especially over Ukraine you know to come to the table when we literally aren't even providing the most basic aspects of of you know our reports and our our key talking points in their languages so these are some of the things where like we just want to be a little bit more forward leaning if only to start great thank you so much to the three of you for sharing all of your insights the experiences that you've had as part of the conflict observatory team over the past year and so I will close by saying thank you Catherine thank you Caitlyn and thank you Andrew for joining us and thank you to the audience for the great questions and engagement and we will close out the panel now thank you