Tracking clues to war crimes
Student researchers at UC lab pore over videos, other evidence of atrocities to aid rights groups
Andrea Trewinnard had been poring over satellite images of the Syrian city of Idlib for hours when she spotted a clue connecting an online video to a bombed marketplace: The shadow of a minaret, indistinguishable to the untrained eye.
“When I see a mosque, I’m happy,” she said.
Trewinnard is engaged in the arduous task of trying to geo-locate and authenticate footage from war zones around the world, making sure videos were shot when and where the people who posted them online say they were. In her study of the Syrian city, the mosque she noticed proved to be a distinctive landmark — corroborating the footage shot at the 2013 marketplace fire caused by a suspected Russian air strike.
The 29-year-old Trewinnard, a first-year law student, is one of a group of UC Berkeley undergraduates, graduate and law students volunteering hours of their free time scrutinizing videos not of the latest viral phenomena, but of air strikes, bombings and craters filled with remnants of weapons.
It’s all part of the Human Rights Investigations
“When I see a mosque, I’m happy.” Andrea Trewinnard, volunteer researcher for the Human Rights Investigations Lab
Lab, an effort to harness the research skills of the budding investigators by having them scrutinize hours of opensource material, looking for evidence that can help groups like Amnesty International raise awareness of global atrocities and prosecute war criminals.
Trewinnard lived in Egypt for three years, and her Arabic background helps her during her searches. In the video of the marketplace fire, a minaret is barely visible through the smoke, but it was enough to send her combing the Internet for mosques next to markets in Idlib and social media posts mentioning the fire.
Her work has paid off in the verification of a number of videos, many of which are part of the Syrian Archive, an effort to preserve media documenting the Syrian civil war so it can be used by future historians, journalists or prosecutors, she said.
“The lab is a way to feel like you’re having some existential contribution to the field,” said Trewinnard. “It’s a way to feel a little less powerless.”
The lab evolved from the Berkeley Human Rights Center’s 2014 Human Rights and Technology program, in which an attempt was made to answer a question vexing advocates fighting for global justice for years: Why were prosecutors at the International Criminal Court — established as a venue to prosecute those who commit genocide and other crimes against humanity — having so little success putting away the masterminds behind high-profile, horrific crimes?
It took the ICC six years after opening its doors in 2003 to hear its first case, and fewer than two dozen cases have made it to the pretrial stage or been brought before the court since then.
Of those, at least five collapsed, like the cases against Kenyan President Uhuru Kenyatta, accused of fomenting a wave of violence after Kenya’s 2007 election in which more than 1,000 people were killed and 900 women were raped, according to the nonprofit Human Rights Watch.
“What our researchers found was that, basically, [prosecutors] were over-relying on witness testimony. While witnesses will always be at the heart of any international prosecution, they’re also extremely vulnerable,” said Alexa Koenig, executive director of the Human Rights Center.
Cases also tend to fall apart because of difficulty in linking “intellectual authors” of the atrocities to those ordered to commit them, said Mayra Feddersen, a doctoral student who works in the Berkeley lab.
Feddersen came to Berkeley after practicing human rights law in her native Chile and growing frustrated with the limited impact she had.
“I realized the law alone is not enough — that only bringing a case to court did not change the reality of many, only for that particular person,” she said.
In comparison, she said, the future of open-source investigations seems full of possibility.
One goal of the lab is to mine social media for evidence of connections between onthe-ground killers and torturers and the higher-ups who direct them.
“Alleged perpetrators are leaving more fingerprints in various places. They’re leaving fingerprints behind the cyber curtain, meaning, they’re going on email, they’re sending messages to their subordinates, they’re filming things,” said Eric Stover, faculty director of the Human Rights Center.
But for materials pulled from YouTube or Twitter to be useful to international prosecutors, they must be determined credible.
“We’re realizing, as there’s a propagation of media and use of media for propagandistic purposes, we need to be really careful about what we’re relying on,” said Koenig.
Human rights organizations rely on trust to do their work, and putting out questionable information can cause that trust to crumble, damaging the whole field, she added.
That’s where the students’ verification work comes in.
By partnering with the students at the Human Rights Investigations Lab, nonprofits like Amnesty International can plow through more material than they would be able to if they had to pay full-time staff to verify videos.
They also have more flexibility to shift from project to project, because the lab can easily assemble teams of students with expertise in a language or region related to a particular project, Koenig said.
Students in the current cohort speak 18 languages, including Burmese, Arabic, Urdu, Hebrew, Turkish, Korean, Bengali and Mandarin. That doesn’t make the work easy, though.
Students scour the Internet for more footage or documentation of the event. They pick apart satellite images. They try to corroborate the videos using details most people wouldn’t give a second thought — plants, clothing, weather, or the minarets of mosques.
“The front-line people who capture images of what’s happening, they tend to focus on the body or where the bomb hit — they don’t realize that’s actually the least helpful info for courts,” said Koenig. “What the courts need is the 360-degree pan shot of the surrounding area to place what happened in a particular location. Or they need information that might seem irrelevant, like the angle something came from, or a shot of a newspaper that establishes the date.”
When Koenig and Lampros launched the lab in 2016, they expected maybe five to 10 students to be interested in the work, which can be tedious and emotionally draining. The students surprised them.
“There was so much demand, I had to cut it off at 42,” said Koenig.
It grew in the next semester to about 60 students. And student interest is also pushing the project to grow in different directions — one group is working with the journalism nonprofit ProPublica to compile a verified national database of hate crimes.
The hope, said Stover, is that Berkeley’s lab can serve as a “blueprint” for other schools to follow.
Some in the legal field are skeptical of the potential value of open-source evidence. The biggest drawbacks: It remains largely untested, can be attacked as unreliable, and has yet to play a key role in a major international criminal trial.
Stover is undeterred. He is not new to exploring untested methods in search of evidence of human rights abuses — like in 1980s Argentina, when he assembled a team of forensics experts to exhume mass graves of thousands of political dissidents, students and other civilians killed during a period of brutal government repression. Argentine archaeologists and scientists didn’t want to participate; some were afraid, and others may have been complicit in the violence being investigated.
“So we had to turn to students at the University of Buenos Aires,” Stover said. “These were 23-year-old students who volunteered, and when we did our second and third exhumations, I even learned one of the students had packed a pistol because he was so afraid. It certainly was new territory.”
For the Berkeley students, knowing their work is of realworld importance helps to keep them going.
“Whenever I’m sitting in front of a computer for five or six hours trying to locate where particular components of a video might be, I try to stop and remind myself, why am I doing this?” said Nisha Srinivasa, 20, a third-year French major at UC Berkeley. “And whenever I stop and remind myself, I realize how exciting the work actually is.”
“The lab is a way to feel like you’re having some existential contribution to the field. It’s a way to feel a little less powerless.” Andrea Trewinnard, volunteer at UC’s Human Rights Investigations Lab