How footage showing the kneecapping of Russian POWs was authenticated
This piece explains the history of open-source investigation of an infamous POW kneecapping video from Ukraine. This investigation substantially consisted of finding and geolocating footage the Ukrainian commander Andriy Yangolenko uploaded to Facebook that related to the liberation of the town where the kneecapping took place. This footage allowed confirmation of Ukrainian responsibility for the abuses.
On March 27, videos began circulating that purported to show Ukrainian soldiers intentionally shooting the knees of Russian POWs. Immediately debate began about the authenticity of the footage, and if it was not fake, about the identities of the POWs and perpetrators. Much of the response was either pro-Russian outrage about Ukrainian “Nazis” being the perpetrators or pro-Ukrainians arguing it was a Russian false flag.
In total, the full video is nearly six minutes in length. The earliest known instances of shorter versions of the video were posted on Telegram starting around 4 A.M. Ukrainian time on March 27, with the almost six-minute version surfacing later. The video was shared widely among pro-Russians, with a major signalboost coming from the pro-Russian commentator Maria Dubovikova on Twitter. Bellingcat founder Eliot Higgins quote-tweeted her, responding that it was a serious incident that deserved to be investigated, and issued an open call for a geolocation as a first step.
Some hours later, I posted the geolocation of the video to a dairy plant in Malaya Rohan, a small village near Kharkiv.
This village had been reported as liberated following fighting on March 25-27. I had begun searching for the location on the assumptions that (1) the video was recently filmed, and (2) it would be near areas Ukraine had recently recaptured. Some of the dialogue featured anger from the soldiers about the shelling of Kharkiv, something Russian artillerymen conducted with brutal effect from positions outside the city - a brutality that had earlier in March elicited the statement from the command of the Special Operations Forces of the Ukrainian Armed Forces that these artillerymen would no longer be captured but slaughtered. Another small village near Kharkiv, Vilkhivka, had also been used to shell Kharkiv city. The Kraken unit had released footage on March 26 that showed Azov forces liberating Vilkhivka. Other videos already circulating from the Vilkhivka operation showed less severe mistreatment of Russian POWs.
This background led me to identify the Kharkiv area as a good bet. The video itself was filmed entirely in an outdoor courtyard with several buildings flanking it. Some trees were evident, and it had an industrial or agricultural aesthetic. Looking for areas matching that description brought me to the Malaya Rohan dairy plant, thus discovering the location of the apparent POW abuse.
In the video itself, the captors seen did not have clearly identifiable insignia, but some had Ukrainian-color armbands. Per professional speech analysis from the BBC, the captors’ voices were in Russian with a Ukrainian accent. The BBC consulted medical and military experts that did not identify anything that pointed to the video being fake, but also not to anything that confirmed the authenticity of the footage.
A major development came on March 27 when Ukrainian journalist Yuriy Butusov uploaded a video that showed the same courtyard as where the torture occurred. The video featured three charred corpses, which he identified as Russian soldiers, and structural damage to the facility not seen in the kneecapping video. The relationship between those charred corpses and the Russian POWs seen earlier, as well as how they came to be burned, remains unknown - although patching together a timeline of March 25 as done below clarifies important details.
Butusov also reported that the Ukrainian combatants involved in the liberation of Malaya Rohan and Vilkhivka were the 92nd Mechanized Brigade, GUR MO (Ukrainian military intelligence), a Kharkiv battalion of Azov, Freikorps, and Slobozhanshchyna. Those other than the 92nd are commonly categorized as “territorial defense units” or “Azov-affiliated.” The perpetrators are presumably drawn from one or multiple of these units.
A leader of the Kharkiv Azov battalion, Konstantin Nemichev, denied involvement in the incident, saying it was “not our location,” possibly a reference to it not happening in Vilkhivka. Two days later, Nemichev and another Kharkiv Azov leader, Sergey “Chili” Velichko (by this point, Russian propagandists and pro-Russian “OSINT” Telegram channels had chosen these two as responsible for the kneecapping), released a statement denouncing the “fake video.”
On March 31, Human Rights Watch issued a statement saying that if confirmed the incident would constitute a war crime and that Ukraine should ensure an effective investigation.
Confirming Ukrainian Custody of Kneecapped POWs
Work to identify the perpetrators did not progress for over a month. On 1 May, I came across a compilation of combat footage that was posted to a small, local channel for Malaya Rohan. There was a scarcity of combat footage from the village, making such a video intriguing.
Upon watching it, a particular detail stuck out: the Ukrainian soldiers took POWs. The potential implications of this were immediately obvious. If prisoners or soldiers could be matched with those at the dairy plant, then not only would the occurrence of a war crime be proved, but it might become possible to name perpetrators. Once I geolocated the footage to less than a kilometer from the dairy plant, I posted a short thread to Twitter on 1 May:
One distantly-seen POW in this combat footage stood out, bearing a resemblance to an orange armband-wearing prisoner in the March 27 video:
At this point, I also shared the latest findings with other open-source researchers based on Twitter, Slack, and Discord. Alexis J Falconcer, an OSINTer mostly based in the Bellingcat community discord, had a premium account for Pimeyes - a facial recognition tool with an imperfect hit rate, but often useful for uncovering leads. Running the face of the most prominent Ukrainian soldier in the compilation got a hit, leading to the Youtube channel of a man named Andriy Yangolenko. His name now identified, he could be searched across social media profiles, leading Alexis Falconer to find his Facebook page.
His FB page featured prolific uploads of combat footage. Starting in late March, Yangolenko had begun uploading clip-show videos showing his unit operating in Mala Rohan, engaging in combat and taking prisoners. His most famous compilation, which I had seen on Telegram, was posted on April 18. It had been crossposted to platforms such as Telegram and TikTok, accruing tens of thousands of views. At this time, a close-up view of the man with the orange armband was not available in any of his videos.
As for Andriy Yangolenko’s identity, his FB videos since February 24 made it clear he was a senior, and probably the senior, leader of a Kharkiv-based battalion called Slobozhanshchyna - a unit Butusov had identified as liberating Malaya Rohan. Simply searching online using Yangolenko’s name in Ukrainian revealed he was a reasonably well-known figure in some circles. He had a history of military service in the Donbass beginning in 2014, a career that included allegations of looting.
In 2015, he was injured in a car bomb that more seriously injured his wife Inna, also an ATO veteran, who would later die in 2020 of COVID-19. Later in 2015, he was detained on grounds of plotting assassinations against public officials, including then-Interior Minister Arsen Avakov. Avakov then personally bailed him out. After February 24, 2022, Yangolenko refounded the 2014-era Slobozhanshchyna, alongside his brother and fellow ATO veteran Serhiy - who was killed during the battle of Malaya Rohan on March 25, along with five other members of the battalion.
On 4 May, Yangolenko uploaded a new video that included previously unused clips. It showed his unit eating and singing together, burying their dead, and, of relevance for this investigation, a clearer video of their capturing the orange armband prisoner. A definite match to the man seen kneecapped could now be made.
What was now established was this: A Ukrainian battalion that helped liberate Malaya Rohan took custody of prisoners later seen being kneecapped less than a kilometer away. Further, the leader of this battalion filmed himself in the same frame as one of these prisoners. While the identity of the Ukrainian soldiers at the dairy plant remain unknown, what is certain is that later-tortured prisoners had come under Ukrainian custody. It is not credible that they could have left Ukrainian custody by the time they were at the plant - meaning the identity of the perpetrators is Ukrainian.
Later Developments
After 5 May, Le Monde began integrating the above findings into a video about the kneecapping. In that process, Le Monde journalist Arthur Carpentier identified two more prisoners who came under Yangolenko’s custody and were also kneecapped. The three men were captured together 750 meters from the dairy plant’s courtyard. Shortly after this footage, they are seen exiting a van that brought them to the dairy plant courtyard and are immediately shot.
Based on historical cloud coverage data for Malaya Rohan, Le Monde concluded that both the video showing Slobozhanshchyna taking custody of the three POWs and the kneecapping video were shot on March 25, not the 26 or 27, and probably less than an hour apart.
After the Le Monde report was published, Yangolenko eventually made a (small) public comment regarding the consequences of his videos. In response to an (extremely biased and factually wrong) Russian media report on Slobozhanshchyna, he posted a short response on his Facebook page, writing:
"SLOBOZHANSCHINA !!!
YOU ARE LIKE STARS IN THE UKRAINIAN SKY !!!
PROUD OF YOU !!!"
[via Google Translate]
Yangolenko has as yet made no further comment.
I initially did not publicize Yangolenko’s FB (and how it was found) out of concern for alerting him to what he had permitted to be proven. I also forensically archived all of the relevant material he provided.
However, contrary to my expectations of an exposed Yangolenko seeking less scrutiny, he did not delete any used videos, or even cease uploads to Facebook. On 1 June, he uploaded footage containing a new clip. In it, he films two bodies lying in a concrete courtyard, easily matched to two POWs seen in the kneecapping footage, and lying in almost identical posture in the same place. It is unclear whether either are still alive. This footage places Yangolenko and Slobozhanshchyna members at the courtyard later that same night.
As far as comments from other Ukrainian battalions who operated in Malaya Rohan, the reaction has been mixed and evolving. When Le Monde contacted Freikorps, they denied involvement, but offered no conclusion as to the video’s authenticity.
For their part, Kharkiv Azov leaders had derided the video as fake or otherwise denied their involvement. However, a video circulated in May among pro-Russian channels that purported to show Chili offering comments to the effect that the kneecapping was Ukrainians’ doing, but that he did not know the full context.
The clip is short and its origin is uncertain. The UN’s Office of the High Commissioner for Human Rights (OHCHR) appears to mention it in a footnote in its June report that includes a section about the kneecapping video, writing that “one of the reported participants of the events later acknowledged that some of his comrades indeed tortured and ill-treated Russian servicemen.” The exact significance of the Chili video is hard to determine without knowing either context or his intentions.
Another important element of reporting has been work identifying the Russian soldiers in the video. The most notable published work on this subject is from Vladimir Sevrinovsky in Meduza, where he reports on a Russian woman’s claim that her son is one of the dairy plant POWs. The man she identifies also appears in Yangolenko’s night video at the dairy plant.
Timeline of March 25-26 Events
Taken together, Yangolenko’s videos and reporting about the town’s liberation allows a rough timeline to be assembled of events surrounding the kneecapping.
A photo that features the Yangolenko brothers can be geolocated to the dairy plant courtyard, and was likely taken on the morning or afternoon of March 25, given the commonalities between it and other footage from March 25, as well as the constraints of Serhiy’s reported time of death and the reported period of battle.
Later on the 25th, Yangolenko filmed the capture of the later-kneecapped POWs. Around the same time, he also filmed himself and his brother interrogating three other POWs (never seen at the dairy plant) elsewhere in town.
This footage can be geolocated to a house along the same central town road, in the opposite direction of the dairy plant.
After the time of the kneecapping video’s filming, once it had become much darker, Yangolenko again filmed himself at this house, with the same POWs still there. This house is roughly 100 meters from where the later-kneecapped POWs were filmed being captured.
It is possible Yangolenko was consistently at this location. It is also possible he or members of Slobozhanshchyna could have been at the dairy plant in the interim, but no visual evidence exists for that conclusion.
Next, Yangolenko filmed himself at the dairy plant standing over the two still bodies of Russian POWs. The plant had not yet suffered the damage seen in Butusov’s footage. Last, Yangolenko’s war videos contain footage shot at night showing debris in the kneecapping courtyard as well as a fire inside the plant to the northwest. This footage is unlikely to have been filmed earlier than the night of March 25, and other information below strongly suggests that is when it was filmed.
This fire is at the same location where destroyed vehicles are seen in photos the Kharkiv regional prosecutor’s office published when they visited the plant in May to document evidence of the Russian use of thermboaric weapons deemed to entail war crimes.
The vehicles are located just outside the kneecapping courtyard.
On May 20, Yangolenko posted a clipshow of Slobozhanshchyna’s visit to the place where he writes that six of their fallen comrades died - which can be geolocated to the dairy plant.
An April 4 FB post by the East Operational Tactical Unit of the National Guard of Ukraine (NGU) provides an explanation of events after the dairy plant came under Ukrainian control.
The NGU account states that the Russian military used incendiary munitions and thermobaric weapons while fighting for Malaya Rohan, sourcing this conclusion from Ukrainian soldiers, destroyed infrastructure, and seized armaments. The weapon system identified as the source of this is the TOS-1A Solntsepyok ("scorching sunlight") system. Relatedly, a TZM-T reloader vehicle for a TOS-1A destroyed during the town’s liberation is well-documented. A child reportedly died from the use of the system on the town.
The report further states that AFU soldiers arrived at “Молочку” (the dairy plant) at night and had to evacuate personnel severely injured from the above mentioned weapons, while themselves witnessing more use of such fire on the plant. A Freikorps soldier interviewed directly identifies the “молокозаводу” [dairy plant] as being hit by incendiary munitions.
The deaths of the Slobozhanshchyna personnel almost certainly resulted from the shelling of the dairy plant, probably from a Russian TOS-1A, late on March 25, after the time of the kneecapping.
The relationship between this shelling and the charred corpses of Butusov’s footage is uncertain. For their part, thermobaric munitions are capable of vaporizing human bodies caught in their second explosion. The extensive damage to the plant also indicates shelling capable of inflicting major structural damage. The shelling of the plant is therefore a credible explanation for the burnt bodies.
Alternatively, there have been proposals that that they were deliberately dosed and set alight. Such theories argue that Ukrainian soldiers might have sought to destroy evidence, further humiliate POWs, or actively killed them in this manner. There is currently no evidence to support those theories.
State of the Investigation
The identity of the Ukrainian servicemembers responsible for the kneecapping war crime remains unknown. The soldiers at the dairy plant could have been a mixed group, sourced from the 92nd and other units. What is clear is that, despite silly uses of OSINT-y language to argue otherwise, some Ukrainian troops perpetrated a war crime.
Ukrainian officials have pledged to investigate this incident as well as other videos that purportedly show Ukrainian war crimes, in contrast to official Russian responses to videos showing apparently Russian forces committing abuses. In this instance, the nature of the abuse as a war crime and the side responsible are both clear.