The Need for an Evidentiary Standard for Open Source Evidence
By Sara Birkenthal
I zoom into the pixelated image of a neighborhood in Bayanoun, a district located in the Northern district of Syria, northwest of Aleppo. Deep in my investigation, I repeatedly watch a video of destruction of this town by airstrikes and use online research tools to attempt to geolocate and verify the video. In the process, I am transported to war-torn Syria. In reality, however, I am safely ensconced in Berkeley, California, at Berkeley Law School’s new Human Rights Investigations Lab.
The Berkeley Human Rights Investigations Lab
Berkeley’s Human Rights Investigations Lab, the world’s first university-based open source investigations lab, launched last semester. Through a partnership with Amnesty International, The Berkeley Lab seeks to bring attention to human rights abuses through human rights reports and journalistic projects. It also seeks to gather evidence of genocide, crimes against humanity, and war crimes for future prosecutions.
The Berkeley Lab is training students to join the Digital Verification Corps by teaching them how to verify hundreds of hours of video footage and photographs of human rights abuses and war crimes from around the world – including Syria, Darfur, and Yemen. Students are also using open source methods to gather evidence of genocide, crimes against humanity, and war crimes for national and international criminal courts. These open source investigators access this information through software, data sets, and tools, as well as legal processes, such as Freedom of Information Act requests.
What is open source investigation?
The work of The Berkeley Lab is in high demand. Human rights investigations increasingly rely on open source intelligence – information obtained from social media and other sources, including YouTube – to chronicle and verify violations of international human rights or humanitarian law. For example, recently the NGO Bellingcat used exclusively open source evidence to document Russian participation in the shooting down of Malaysia Airlines Flight 17 over Ukraine.
Benefits and Challenges
The Lab, and similar open source investigations labs established at the University of Essex and University of Pretoria have a number of other benefits: they were created with minimal startup costs and draw on the diverse expertise of these universities’ faculty and students. Notably, since the Berkeley Lab opened its doors last semester, more than 100 students from across the university have joined. The open source movement has opened the door to students who might not have otherwise undertake human rights research, bringing new perspectives and talents to the table.
However, the open source movement is not without its individual and macro-level challenges. One of the biggest challenges for open source investigators themselves is verification of media that they find online or that they receive. Bellingcat’s founder, Eliot Higgins, a Research Fellow at the Berkeley Human Rights Center, has explained the importance of geolocation to verify that an image is what it’s claimed to be. Geolocation is the process of using photos and videos to find the precise location – ideally, down to the latitude and longitude coordinates – of what happened. The geolocation process can often take hours for a short clip, and can involve combing Google Maps and other online resources for landmarks that stand out amongst the rubble. Once an investigator can make a connection between a neighborhood where he or she suspects an event has taken place and a landmark in that neighborhood, he or she can go about attempting to confirm the coordinates of the location. However, the investigation process is inherently uncertain; an investigator often finds multiple landmarks in a neighborhood that could be the site of an event. Investigators learn the importance of pursuing numerous avenues and not allow bias to lead them to conclude that something is a certainty, when, in fact, it is merely a possibility.
On a macro level, while the open source movement seems promising, the current reality is that the question of whether the open source data will be used as evidence in future tribunals is still an open one. Open source evidence remains largely untested in international tribunals due to the uncertain evidentiary status of open source materials. A recent example from the International Criminal Court (ICC) illustrates the uncertainty of the future of open source evidence. The ICC indicted former Congolese Vice President Bamba for murder, rape, pillaging, and bribery. The prosecution submitted evidence showing a wire transfer from Bamba’s sister to a witness, who allegedly passed the money to another witness, who in turn purportedly gave false testimony. The prosecution also submitted Facebook pictures reportedly showing the two corrupt witnesses together. In response, the defense disputed whether these images had probative value, as required by the court’s governing law, the Rome Statute, and its Rules of Procedure and Evidence. According to the defense, it was impossible to know who posted the videos, when and where they were taken, who took them, and if the people in them are who the prosecution claims they are. Further, the defense found problematic the means by which the prosecution obtained the photos: extracting them from Facebook. Because the prosecution does not have access to metadata – such as the IP address of the uploader or the time stamp – there is no way to determine who posted the photos.
To date, it remains to be seen how the ICC will come down on this issue. The court’s flexible evidentiary standard, which does not designate categories of inadmissible evidence, but rather establishes a paradigm for analysis of evidence, makes it hard to predict how the court will rule on the prosecution’s submission. This lack of clarity has the effect of disincentivizing investigators and parties from utilizing open source evidence. If parties submitting open source evidence – much of which is new to the court – cannot rely on a particular determination from the court as to the admissibility of that evidence, they will be less likely to submit such evidence. Accordingly, in its place, investigators and parties will continue to depend upon traditional methods of investigation, which can be time-consuming and costly in terms of the burdens undertaken by witnesses, many of whom risk their lives to testify against powerful individuals whose resources may exceed the operating budget of the tribunal. Even recognizing that under the court’s current flexible evidentiary standard, it will likely admit the photos in the Bamba case, and similar open source evidence, the opportunity cost in time spent waiting for the court to render a decision on its admissibility nevertheless slows proceedings by a court widely criticized for being inefficient.
Though the use of open source evidence in investigations certainly has its challenges, given the widespread adoption of Internet-connected mobile devices and social media, failing to use open source methods to research and verify human rights abuses and war crimes is an enormous missed opportunity. International courts, such as the ICC and other tribunals, should encourage the use of open source evidence by replacing the current general standard with one specially-tailored for open source evidence.
Like traditional types of evidence, open source evidence should not be assessed in a vacuum; instead, according to the court’s admissibility test, open source evidence admissibility should be determined as part of the evaluation of probative value, taking into account relevance and prejudicial effect. Further, an evidentiary standard for open source evidence should safeguard open source evidence from tampering, tainting, and corruption. Importantly, an open source standard should differ from the existing general evidentiary standard in that the current standard bars UN or NGO reports that do not provide sufficient detail about their sources; an effective open source standard would have to be more lenient to take into account the fact that open source evidence, by its very nature, is often anonymous. According to this approach, open source evidence could be admitted even when its origins are unknown, but the weight afforded to the evidence may vary.
Considering the advantageous nature of open source evidence to the ICC and other tribunals, investigators should have an easily-accessible framework to evaluate open source evidence. While the importance of an evidentiary standard for open source evidence is clear, little progress has been made thus far to develop and implement one. As open source evidence inevitability proliferates in the years to come, a tailored standard will become increasingly critical, and lack thereof will become increasingly grave for international courts and the victims they are intended to protect.