7. Using UGC in human rights and war crimes investigations
Written by Christoph Koettl
In the early summer of 2014, Amnesty International received a video depicting Nigerian soldiers slitting the throats of suspected Boko Haram supporters, and then dumping them into a mass grave. The video, which circulated widely in the region and on YouTube, implicated Nigerian soldiers in a war crime. However, in order to draw that conclusion, we undertook an extensive investigation involving video analysis and field research. This resulted in the publication of Amnesty International’s (AI) findings of this incident.
This incident is a powerful example of how user-generated content can contribute to in-depth investigations. It also demonstrates the importance of digging deeper and going beyond the basic facts gathered from standard UGC verification. This is particularly important for human rights investigations. UGC not only aids in determining the place and time of a violation; it can also help with identifying responsible individuals or units (linkage evidence) that can establish command responsibility, or with providing crucial crime base evidence that proves the commission of a crime.
While there are differences between human rights and war crimes investigations and journalistic reporting, there is also immense overlap, both in regards to the verification tools used and in terms of the benefits of relying on UGC. In fact, the British media outlet Channel 4 conducted an investigation into the conflict in northeastern Nigeria that was largely built on the same UGC footage.
Principles of human rights investigations
While a lot of UGC might have immense news value, human rights groups are of course primarily interested in its probative value. In a human rights investigation, we compare all facts gathered with relevant human rights norms and laws (such as human rights and humanitarian, refugee and criminal law) to make determinations of violations or abuses. Consequently, a single analyst who looks at UGC, such as myself, must be part of a team comprising relevant country, policy and legal experts.
Our ultimate goal is to achieve a positive human rights impact, such as when our work contributes to establishing an international inquiry, or the indictment of a suspected perpetrator. Today we are achieving the best results when combining a variety of evidence, such as testimony, official documents, satellite imagery and UGC.
This requires the close collaboration of researchers who possess country expertise, trusted contacts on the ground, and highly specialized analysts who do not focus on a specific region or country, but are able to provide analysis based on satellite imagery or UGC.
In some instances, one piece of evidence does not corroborate some of the information gathered during the investigation, such as when satellite imagery does not support eyewitness claims of a large mass grave. We then exercise caution and hold back on making statements of fact or determinations of violations.
This close collaboration among a range of experts becomes even more relevant when going beyond war crime investigations, which can be based on a single incident caught on camera. Crimes against humanity, for example, are characterized by a systematic and widespread nature that is part of a state or organizational policy. Research solely based on UGC will hardly be able to make such a complex (legal) determination. It usually provides only a snapshot of a specific incident. However, it can still play a crucial role in the investigation, as the following example will show.
War crimes on camera
In 2014, AI reviewed dozens of videos and images stemming from the escalating conflict in northeastern Nigeria. Human rights groups and news organizations have extensively documented abuses by Boko Haram in the country. But this content proved especially interesting, as the majority of it depicts violations by Nigerian armed forces and the state-sponsored militia Civilian Joint Task Force (CJTF).
The most relevant content related to events March 14, 2014, when Boko Haram attacked the Giwa military barracks in Maiduguri, the state capital of Borno state. The attack was captured on camera and shared on YouTube by Boko Haram for propaganda purposes. It resulted in the escape of several hundred detainees. The response by authorities can only be described as shocking: Within hours, Nigerian armed forces and the CJTF extra-judicially executed more than 600 people, mostly recaptured detainees, often in plain sight, and often on camera.
Thorough research over several months allowed us to connect different video and photographs to paint a disturbing picture of the behavior of Nigerian armed forces. For example, one grainy cellphone video showed a soldier dragging an unarmed man into the middle of a street and executing him, next to a pile of corpses.
We first performed standard content analysis. This involved extracting the specifications of the road and street lamps, buildings and vegetation, as well as details related to the people seen in the video, such as clothes and military equipment. Reviewing the video frame by frame greatly aided with this process. The geographic features were then compared to satellite images of the area on Google Earth.
Based on this work, it was possible to pinpoint the likely location within Maiduguri, a large city of around a million people.
Several months later, additional photographs, both open source and directly collected from local sources, were used to paint a more comprehensive and even more worrisome picture of the incident. For example, at least two of the victims had their hands tied behind their backs. It is noteworthy that several photographs in our possession were actually geotagged. We discovered this by using a EXIf reader to examine the metadata in the photo. This location data proved a perfect match to the street corner we identified as part of the content analysis of the initial video.
Other videos from the same day documented an even more gruesome scene, which suggested another war crime. They show the killing of several unarmed men, as detailed earlier in this chapter. The videos were a textbook example of how UGC can be a powerful tool in longterm investigations when combined with traditional investigative methods.
We slowed the video to perform a content analysis in order to identify distinctive markings on the soldiers and victims, or anything that could indicate location, time or date. This revealed two important details: a soldier wearing a black flak jacket stating “Borno State. Operation Flush,” the name of the military operation in northeastern Nigeria; and, for a split second, an ID number on a rifle (“81BN/SP/407”) became visible. No distinctive geographic features were visible that could be used to identify the exact location.
AI subsequently interviewed several military sources who independently confirmed the incident, including the date and general location outside of Maiduguri. An AI researcher was also able to secure the actual video files while on a field mission to the area. This allowed us to conduct metadata analysis that is often not possible with online content, since social media sites regularly modify or remove metadata during the upload process.
The data corroborated that the footage had been created March 14, 2014. Obtaining the original files is often possible only through well-established local contacts and networks, who might share content in person or via email (ideally encrypted). Savvy news desk researchers and journalists who might be inclined to contact local sources via Twitter or other public platform should consider the risk implications for asking for such sensitive footage from contacts in insecure environments.
In this case, two sources stated that the perpetrators may be part of the 81 Battalion, which operates in Borno state, and that the rifle ID number refers to a “Support Company” of that battalion. Most important, several sources, who had to remain anonymous, separately stated that this specific rifle had not been reported stolen, disqualifying the predictable response by Nigerian authorities that the soldiers were actually impostors using stolen equipment.
After an initial public statement about the most dramatic footage, AI continued its investigation for several months, bringing together traditional research, such as testimony, with satellite imagery and the video footage and photographs detailed above. This UGC supported the overall conclusion of the investigation that both Boko Haram and Nigerian armed forces were also implicated in crimes against humanity. These findings can have serious implications, as the violations detailed are crimes under international law, and are therefore subject to universal jurisdiction and fall under the jurisdiction of the International Criminal Court.