An official website of the United States Government Here's how you know

Official websites use .gov

A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS

A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.


  • In this on-the-record, virtual briefing, research experts from the Yale Humanitarian Lab (Yale HRL) discuss findings in a report supported by the U.S. State Department that details the detentions and disappearances of 226 persons in Kherson, Ukraine, as they relate to Russia’s ongoing filtration operations. 


MODERATOR:  Okay.  So hello and welcome to the Washington Foreign Press Center’s briefing on accountability for war crimes and other atrocities in Ukraine, and the unjust detentions and disappearances in Kherson Oblast.  My name is Doris Robinson and I’m the moderator for today’s briefing.  Joining us today is Nathaniel Raymond from Yale University and Caitlin Howarth, director of operations for the Conflict Observatory team based at the Yale Humanitarian Research Lab. 

Now for the ground rules.  This briefing is on the record.  We will post the transcript and video of this briefing later today on our website at  Please make sure that your Zoom profile has your name and media outlet that you represent.  Each of our briefings will give opening remarks and then we will open it up for questions.  We will start with Professor Raymond followed by Caitlin Howarth. 

With that, over to you, Professor Raymond. 

MR RAYMOND:  Thank you so much, Doris, and it’s always a pleasure to be here at the FPC.  And thank you all for joining.   

Today we are discussing a recently released report about disappearances and detentions in Kherson Oblast in Ukraine.  This report is really the result of several weeks of intensive open-source data analysis and supporting imagery analysis by the team here at the Yale Humanitarian Research Lab as part of the Conflict Observatory, and it is – we think all of our reports are important, but this is, I think, our – one of our most important reports to date because it provides, as you will see, confirmation about the fears expressed prior to Russia’s invasion of Ukraine in February of this year that Russia and its military and security services would use targeted lists to capture, detain, and in some cases allegedly kill members of civil society, government officials, and members of vulnerable populations such as ethnic groups.   

And can we start the slides?  Thank you.  Next slide, please. 

So before we get into the findings of the report and the methodology by which we arrived at those findings, the headline here is that this report documents multiple alleged violations of international law by Russia’s military and security services, in particular the FSB.  These violations include violations of the Geneva Convention, violations of the Convention against Torture, violations of the UDHR, and, most notably, where those rights are enshrined in part in the International Covenant on Civil and Political Rights, particularly as it relates to torture; cruel, inhuman, and degrading treatment; enforced disappearances; detention; threats to relatives and loved ones, including threat of torture, detention for refusal to vote in the annexation referenda; et cetera.  These violations also include alleged violations of the Rome Statute, particularly Article 7 and Article 8.   

Next slide. 

So where we want to begin today is in February 2022, when U.S. Ambassador Bathsheba Crocker wrote to the then High Commissioner for Human Rights Michelle Bachelet warning that Russia may engage in operations that – and I’m going to read the bolded part of the letter here – that include “targeted killings, kidnappings, forced disappearances, unjust detentions, and the use of torture, would likely target those who oppose Russian actions.”  And in the italics you see some of the potential populations that those operations would target: journalists, anticorruption activists, and vulnerable populations such as religious and ethnic minorities and LGBTQI persons.  So Ambassador Crocker states that we – the State Department – “have credible information that indicates Russian forces are creating lists of identified Ukrainians to be killed or sent to camps following a military occupation.” 

Why we’re starting here is that the evidence we are going to present to you today suggests that this is exactly what the Russians and their military intelligence services have been doing in Kherson Oblast and likely in other areas of Ukraine.   

Over.  Next slide. 

So these alleged use of targeted lists suggests that – and our evidence shows that – these lists were prepared likely before the invasion but have been added to and edited afterwards, that they use social security records, that they use information from school records and principals and information about residents from building managers in apartment complexes.  There’s also evidence that Russia’s forces have been using photographs of attendees at protests and that they’ve been targeting veterans, specifically of fighting in Donbas, members of the law enforcement community, emergency services personnel.  And they’ve also been targeting Crimean Tatar leaders, especially affiliates of local Tatar groups.  And they have been targeting protesters and online activists.  You see the quote here:  “It was a planned purge of the population.  They took everyone who could in any way resist.”   

Next slide.   

I’m going to turn it over in a second to my colleague Caitlin Howarth, but first I want to really be specific about how we’re defining detention and disappearances.  So we’ve documented, through the use of open-source information, including social media, including news reports, including statements by Russia itself, approximately 226 individuals in Kherson Oblast since Russia initially took control that have been disappeared or detained.  The evidence is consistent with an intentional campaign of detention and disappearance.  And this is one of the most disturbing findings, is that one-quarter of the individuals in this data set report what can constitute torture or cruel, inhuman, and degrading treatment.  And at least five deaths in custody or soon after being in custody appear to have allegedly occurred.   

Unlike the Donetsk filtration system we documented in a previous report where proxy forces of the Donetsk People’s Republic were deeply engaged in those operations, these operations in Kherson seemed to have been largely led and executed by Russia’s military and the FSB.  Approximately half of those detained or disappeared in our data set do not appear to have been released as of the fall or just before the fall of Kherson.   

This is an important point, the last one, and for me it’s one of the most critical findings of the research, is that approximately half of this data set includes individuals who were taken at their home.  That’s really a critical piece of corroboration about the use of targeted lists, because to target people at their home you have to know who you’re targeting and where they are.   

What’s also interesting here in this last finding is that members of the Crimean Tatar population appear to be disproportionately targeted at checkpoints compared to non-Tatar community members.  And this disparity or difference between how Tatar and non-Tatar community members who’ve been targeted and detained and where they’ve been targeted is one of the most, I think, critical findings of the report.   

Next slide.   

So with that, I’m going to hand over to Caitlin, who’s going to walk us through the methodology.  Then we’re going to go deeper into some of the specific findings, particularly on torture and deaths in custody.  Caitlin.

MS HOWARTH: Thank you so much, Nathaniel. 

One quick point on the findings that Nathaniel just offered to you is that our reporting was predominately drawn up to the date that the report was issued.  Most of our material was drawn prior to the liberation of Kherson.  So prior – so a lot of our material – we were drawing on reports immediately prior to the release of the report.  However, given that fact, it’s important to understand that these are likely significantly undercounting other new facts that are likely now going to be present on the ground.  So you should be on the lookout for some upcoming reports from us that will significantly update these numbers and will give new details into what has become of people who, as of now, we can only document up to the point of their disappearance or detention.  

On to our methodology, how we actually were able to identify individuals, identify detention locations, and the challenges and limitations of doing all of this.  When it came to individuals, it’s important to understand that the individual accounts recorded are gathered from a variety of different locations.  Social media posts – these are user-generated documents that are posted by the individuals themselves and in many cases by families or eyewitnesses to their detention or abduction, other instances related to how they were detained or just – or potentially disappeared.  In some cases we also were able to document individual accounts based on the statements of governments of either Ukraine or Russia itself.  In other instances we were able to document individual cases based off of news media reports that were generated on their behalf.  Some of these were generated by human rights organizations.  I should note that in the case of Crimean Tatars, human rights organizations played a particularly helpful role, and publications of the governments of Ukraine and Russia were also specifically important in these cases.   


It’s important to keep in mind that the individual account of a single person on its own does not meet a high confidence standard.  We have to work according to the Berkeley standards and those protocols are very specific in terms of building credibility over a number of sources and from multiple angles.  So any single source cannot by itself reach high confidence alone; or, if it does, it usually has to include multiple things that can then go on to be independently verified. 

In order to build that level of high confidence, we had to be able to combine additional allegations into the aggregate.  So it’s important to kind of keep in mind that as we’re building additional accounts we have to look at whether we can build credible allegations in the aggregate in order to achieve a high confidence standard. 

On detention locations, these were also built from both primary and secondary open sources, and in all of these cases we were able to achieve a primary geolocation verification through satellite imagery, and we were only able to achieve that high confidence standard once we had found five high quality and independent sources which, in this case, could have included also three additional individual narratives identifying that site as a location connected to the detention of persons. 

Some of the challenges related to this, as I mentioned before, we had some specific challenges related to individual accounts and how many independent points of verification we could find based off of a single account and getting to a comprehensive level.  We definitely did not have a – either a comprehensive nor a random sample.  When working on instances of detention and disappearance, this is something where one is not going to necessarily achieve a random sample.  We have an ethical obligation to document every case that we can find, but that by definition also means that we are not achieving a random sampling of an entire population. 

When reporting – when gathering reporting, there is also certain biases that are introduced in terms of who is reported on, which cases are considered to be relevant which are reported on in this kind of context versus cases that might be deemed missing but not necessarily related to an occupation. There are cases that are considered also to have certain survivorship and release biases.  Not everybody’s individual case is necessarily reported on or discussed over social media or discussed with a reporter or a human rights organization, and that may be based off of the understanding that a family or bystanders have in terms of the sensitivities related to that specific case. 

And satellite imagery, of course, also this is imagery that cannot view incidents or activity that are taking place inside specific buildings, so there are limitations to what we are able to confirm using satellite imagery alone.  That is why we make sure to augment this with additional open source images, photographs, video, other things that are taken from different vantage points.  That is also why we expect that we’ll be able to augment this particular report with new reporting that will come based off of additional research in the coming months.   

On to the next slide, please.   

Some interesting trends that we found when trying to get a better understanding of who has been detained over the last several months.  First, you already heard a bit of our concerns about the Crimean Tatar population.  In Tatar-majority towns, we have seen that these have appeared to have received some more attention, and those who have been accused of membership in a Tatar voluntary organization were disproportionately likely to be detained.  They were also disproportionately likely to be charged with a crime.   

Now, this is important because there are actually – the clear majority of those who were detained or who seem to have been – who seem to have suffered an enforced disappearance do not appear to have been charged with any specific legal crime or to have entered into the criminal justice system in any way according to our data so far.  However, in the case of Crimean Tatars, we do see that they have actually been charged with crimes.   

And so this is an important distinction.  We do, I want to note, see that there are more details related to these cases, and in part that is due to the fact that they have been documented somewhat more comprehensively due to some of the excellent reporting and follow-up that’s happened from Crimean Tatar human rights organizations.  So there may be some discrepancy introduced from that, but we want to emphasize as well that there are some other instances here where maybe some of the – some of the details may also just be some discrepancies.  So it’s something that we continue to watch.  

Of those who are detained, the vast majority are male; and of those, the clear majority were of military age.  A very significant number were government officials.  Another significant number were civil society or volunteer, and volunteer has a very specific meaning in this context so I would urge you all to look at the report and see some of the definitions that are used in this context.  Teachers also appear to have been an important target group on this list, as were journalists.  Among those who were detained, there were a half dozen instances of family members than subsequently being detained.   

And when it came to electronic surveillance and monitoring, we saw that about 24 individuals had personal electronics taken, accessed, or monitored.  And when it came to monitoring, we do see that there were clear instances of monitoring being reported by family members or others associated with those detainees.   

So these would be instances where, for example, a detainee having been taken and known to be in detention, a family member or somebody else who they knew would attempt to contact them over their digital device, would see that their message had been read, had been accessed, and in some cases would then get a response to a message or would have some form of outreach, but that the response would be in a manner inconsistent with their loved one.  And this was significantly higher rate among Crimean Tatars where the rate of electronics taken, accessed, or monitored was all the way up to 19 percent.  And again, I want to stress that those are confirmed instances, and the actual rates may have been higher.   

Finally, it’s important, again, to note – I’m sorry, can we move on to the next slide, thank you – that among the forces engaged in detention and disappearance here, you’ll see a number of different sort of groups here.  Russia’s military formed by far the clearest group.  FSB were also among this group, and you’ll also see a note here for election workers.  These are people who were accompanied by armed forces going door to door during the recent referendum.   

There is an important discrepancy here in that, as opposed to Donetsk, there was very little observed involvement of any sort of proxy forces.  So instead of seeing proxies for Russia’s military or other sort of non-official or associated forces with Russia’s armed forces, in these cases we saw clearly that these were Russia’s military or FSB in the vast majority of cases.  There were some unknowns that were also documented, but in this case truly a clear majority.  And these were also not cases where somebody would simply say, yes, it was Russia’s military and then move on.  These were documented instances where emblems were observed, uniforms were observed, there’s additional documentation to demonstrate that in fact these are correctly identified as military.   

It’s also important to see some of the cases in which the FSB were specifically involved, so I urge you to look – for those that are interested in those details, I do urge you to look into the report more fully to see where they appear specifically.   

Next slide.   

Here, I’m going to hand it back to my colleague Nathaniel to explain in somewhat more detail the significance of the kinds of abuses that were observed and documented in this report.  Nathaniel, back over to you.   

MR RAYMOND:  So thank you, Caitlin.  I have been working on documenting torture cases now for about 23 years, and one thing I’ve learned during that time is that those who are subjected to torture usually under-report what they have been through, especially in cases involving sexual gender-based violence and psychological torture and humiliation.  What – why I’m saying that is that experience, for those of us who have been documenting torture cases, makes these numbers shocking that we have 55 people amongst the data set alleging torture in very commonly repeating themes in terms of the tactics used: stress positions, beatings, electric shock, sexual gender-based violence which includes in six cases individuals being subjected to SGBV or threatened with it, including beating of – intentional beating of genitals.   

The – what’s also stands out here is the use of mock executions, including Russian roulette on multiple detainees and the use of sleep deprivation, sensory overload, temperature manipulation, which are commonly seen in terms of interrogation settings and have been known to be used by the FSB in the past.  There was a clear theme of conditions of confinement that rose to the level of alleged cruel, inhuman, and degrading treatment, including a lack of medical care for those who’d been injured – including injured during confinement – overcrowding, insufficient food and water – and that propaganda videos also were seen, that 12 individuals reported being forced to record some form of a propaganda video, which is a violation of international law.   

The last point on this slide is the first bullet.  At least four deaths in custody were reported, including one immediately following of an individual who had been released who had been allegedly subjected to techniques tantamount to torture.   

Next slide.   

So an introductory point here to the detention checkpoint locations.  We identified, as you see on the side, 12 locations with high confidence.  For reasons of operational security, even though Kherson fell after the report was released, we are not identifying the specific locations and names for a variety of reasons.  One, we do not know where all of the alleged detainees in this data set were or are.  Additionally, we want to make sure that evidence at those sites is held by the proper authorities with the right chain of custody procedure.  So these locations have been shared with those who need to have them.  And because we’re not listing these names and specific locations, it doesn’t mean we don’t have them; we do.   

In terms of checkpoints – and you’ll see some imagery examples here in a moment – the checkpoints were extremely common.  And we – as we mentioned before, they appear to be the locus point for the detention of – or of the initial capture of Tatar members of the community.   

Next slide.   

So here’s an example of a checkpoint site seen in satellite imagery provided by the Maxar corporation.  You can see a queue of cars waiting to enter a checkpoint area.  There’s a convoy of Ural-4320 military vehicles.  There is a covered area, where it appears that the cars are being processed through.   

Next slide.   

This is an example of detention locations that we saw, and we were able to geolocate them through non-image resources and then corroborate that with image resources, including the following indicators: presence of military vehicles such as these Kamaz and Ural-type trucks, and also road obstacles or checkpointing around the facility.   

Next slide.   

MS HOWARTH:  I believe that’s our final slide.   

MR RAYMOND:  So no more slides.  Thank you.  (Laughter.)   

What I want to end with here is just reiterating the three top headlines.  One, this is corroboration of concerns expressed by the U.S. Government prior to Russia’s invasion that targeted capture-and-kill lists would likely be used by Russia and its military and intelligence services.  Second, that there is clear evidence of targeting of Crimean Tatars and of targeting of civil society members through the use of both snatch-and-grab at checkpoints and house-to-house abductions.  Third and finally, there is clear evidence of deaths in custody and the widespread use of techniques widely recognized as constituting torture and cruel, inhuman, and degrading treatment, including mock executions, SGBV, beatings, stress positions, electric shock.  And this resulted in at least as many as five deaths – four of them in custody and one soon after.   

The last point to end on here before we go to questions is that – and it really is probably the most telling part of our research – the 226 individuals in this data set, why 226?  Well, we could have kept going; there were so many cases.  But at some point, we had to stop and present the evidence we had.  So, as Caitlin mentioned, this is not a representative sample, and unfortunately it is probably only a fraction of the available incidents of individuals being detained, allegedly tortured, and in some cases disappeared. 

So in closing, this is not just an accountability issue in terms of alleged war crimes and potentially crimes against humanity.  It is also a humanitarian issue and a humanitarian emergency in terms of identifying the remains of any individuals who died in this system, of working with loved ones who are desperately trying to find out where their family members are, and of providing psychosocial support and health care to survivors of this system that appears to be relying on the widespread use of torture. 

Over to you, Doris, and to our colleagues for questions. 

MODERATOR:  Great.  Thank you so much for those remarks.  We’ll go ahead and start our question and answer session.  For the journalists, if you have a question, please go to the participant field and virtually raise your hand.  I’ll call on you, and you can unmute yourself and ask your question.  And I see we have our first question from Alex Raufoglu from Turan News Agency. 

QUESTION:  Doris, thank you so much.  And I thank you both so much for your compelling report, and for making yourselves available for us today.  So much to digest here, and sadly I have to drop off for another briefing in a few.  But let me try to follow up with a few details, if I may.  Nathaniel, how much do you think your finding is consistent with early reports on FSB, quote/unquote, “kill list” – something we have discussed before, I remember, earlier in the year?  And as you know, some of them have been even published before the invasion. 

For Caitlin, you mentioned that they also use some social media tools.  Can you please expand on that?  What tools have they turned to mostly?  Twitter, Facebook, WhatsApp?  And is it your impression that they use those tools for locating their targets only, or for somewhat, like, comprehensive researchers which might or might not require access to additional services, such as paid services?   

And lastly, if I may, to you both:  As you made it clear that putting reporters, journalists, critics in – on kill lists is a violation of international law, while this might offer to Western audience a sort of new angle to approach Russian war crimes, but for most of us who have been covering Russia’s domestic politics as well as its crimes in many parts of the world – we have Syria and other examples for how far Russians can go on this – this is not completely new.  So do you think the fact that this remained underreported and in many cases unaddressed in other regions, as well as inside Russia of course, is the reason why we see what we see today in Ukraine?  Thank you so much. 

MR RAYMOND:  Great questions as always, Alex, and good to see you.  I’ll take question one and three, and then turn it over to Caitlin for your second question about social media.  First, what we are seeing here is consistent with FSB activity, both inside Russia and elsewhere outside Russia.  Sadly, there was nothing that we found that was particularly surprising for those who have studied the methods and record of the FSB in terms of their role in detention and interrogation. 

To your third question, I can answer it very directly and simply.  Impunity for past crimes is the most effective recipe for the commital of future crimes.  The culture of impunity that has existed in Russia in the Putin regime and the actions of security services on Russia’s people show clearly that the behavior in Kherson is not an anomaly, it is simply an import into the Kherson Oblast of alleged methods used within Russia by security services there on their own people.  And again to stress, impunity for crimes in the past will lead to crimes in the – today and the future. 

Caitlin, over to you for question two. 

MS HOWARTH:  Certainly.  So this will probably be a frustratingly broad response, but our methods are essentially we will – we’ll collect from as many sites as are relevant when it comes to social media.  Anything that is user-generated media is going to be relevant to our collection.  That will mean in these cases anywhere that potential families, detainees, anywhere that folks who are – who have information that may be relevant to any given investigation are posting is where we will be – is where we will attempt to find that.  So it’s always going to be important for us to be looking more broadly and making sure that our methods are specific to any given investigation, and for that we use a variety of tools.  We try to make sure that those tools are, again, specific to the different channels, but it can be surprisingly low-tech at times.  So it’s always something where we have to make sure that our team works with the kind of – with the kind of security that they need to be able to move and be as safe as possible while they conduct their research.  So I hope you understand that that’s always a moving target on any given day.  Thank you.   

MODERATOR:  Thank you.  We’ll take our next question from Mathias Ask, TV2 Norway. 

QUESTION:  Hi.  Can you guys hear me?   

MR RAYMOND:  Yeah.  Hi, there.   

QUESTION:  Okay.  Hi, thank you for this briefing.  I just have a question about the timeframe because I think this report covers March to October, and I was wondering, now with Kherson being liberated by Ukrainian forces so recently, do you anticipate doing more research into this region?  And do you think there’s going to be more available evidence considering now that the Russian troops are no longer there? 

MR RAYMOND:  At this point, we are for the moment on other priorities, which hopefully we will be able to share the output of that with you soon.  In terms of looking at our work on Donetsk filtration system as a precedent, once we do an investigation we constantly continue to revisit it and look for additional information that allow us to both confirm or refute our previous findings and expand upon them.   

So at this point, while we do not have anything new on Kherson, we assume with past liberated areas as precedent that we will have more.  And when we do have more, we will report it.  At this point, what we are most interested in is building on the Kherson work into any other recently liberated areas or areas that may be liberated in the future to see if we can detect patterns consistent with what we have seen either in Kherson or Donetsk.  So we are still on the case.  For the moment, we have shifted to another priority, but we have a watching brief on Kherson and similar activities.  So the answer is yes.  Over.  

MODERATOR:  Great.  Thank you.  I will do one more round of questions.  If you have a question, please raise your hand virtually.  You can also type your question in the chat box.   

And it doesn’t appear we have any more questions.  So Nathaniel, I will turn it back to you for any closing remarks. 

MR RAYMOND:  Well, I think the only closing remark from my side – and then, Caitlin, if there’s anything you’d like to add – is really directed at all of you in the media.  The work you are doing in following these stories is critical, not just for history and not just for accountability but for the families of those who are looking for answers and are looking for witness.   

So what I would say is continue to tell these stories about the missing, because while the cameras will shift on to the next crisis both within Ukraine and elsewhere, the stories – as someone who has worked with missing families and on missing persons for a long time – these stories take years to tell.  Our report is simply the beginning of the story, but it’s not the end.  And so I would just call on all of you to continue telling the stories of the missing in Kherson and elsewhere, because these stories take time and commitment, but they matter.   

Caitlin, anything to add? 

MS HOWARTH:  Only to echo what Nathaniel just said and to add that some of the most important details that will come out ultimately about this conflict will be with those who have been missing, with those who’ve been tortured, and who eventually, with time, tell more of their story.  They will – sometimes time helps, and sometimes it will also help to uncover some of the most critical details and trends.  So your steadfastness will be really important to helping us to understand better of what actually happened.  We don’t have the whole story yet, not by far.   

Thank you so much for your time. 

MODERATOR:  Thank you both.  I would like to thank our briefers today, Nathaniel Raymond and Caitlin Howarth, for taking the time to brief with us today.  And this concludes today’s briefing.  Thank you all. 

U.S. Department of State

The Lessons of 1989: Freedom and Our Future