It’s taken 54 days for the first findings from the investigation into the crash of Malaysia Airlines Flight 17 from Amsterdam to Kuala Lumpur to be released. The preliminary report by the Dutch Safety Board features several key points: “high-energy objects penetrated the aircraft from outside”; the plane broke up in mid-air; the pilots in the cockpit were talking normally until the plane was hit, indicating they were taken by surprise; the other black box, which recorded telemetric data, also showed normal data until the flight was suddenly hit.
This is all consistent with the Ukrainian government’s belief that pro-Russian separatists shot MH17 down with a Soviet-era Buk ground-to-air missile launcher, and it contradicts the Russian government’s insinuation that a Ukrainian figher jet was detected “in close proximity” to the plane just before it crashed. It’s also consistent with the work of users of a website called bellingcat, who have been cataloguing and analysing photographs of the MH17 crash site for almost every one of those 54 days. The consensus among them is that MH17 was hit by a Buk launcher with the missile exploding on the front port side of the plane, and the shrapnel causing extensive structural damage and near-instant decompression.
Anyone who disagrees with this assessment is welcome to try and debunk it – the possibility of being wrong is a key part of how bellingcat works. It is, in the words of founder Eliot Higgins – better know by his pen name Brown Moses – an “open source investigation”.
A typical verification thread on bellingcat’s Checkdesk site, with users discussing evidence of damage to the nose and cockpit of MH17.
For those who pay attention to the current conflicts in the Middle East, Brown Moses is likely a familiar name. Working from a laptop in his living room in Leicester, he has produced some of the most important journalism on the Syrian Civil War, and become a lodestar for an emerging kind of online citizen war reporting. His first big scoop came after the chemical gas attack on the Damascus suburb of Ghouta on 21 August 2012, where he used video footage posted online to build a case that the Assad regime was responsible for the deaths of as many as 1,729 people. He cross-referenced the shapes of the canisters which had held the gas with those seen in earlier YouTube videos showing other attacks on rebel positions by regime forces – it was the kind of detail only he, having spent hours obsessively gathering such videos, would recognise.
He has no background in weapons or war – he used to work in finance – but he does have a fastidious, obsessive thoroughness about his work that has led New York Times war correspondent C J Chivers to call him “an indisputable resource”. Human Rights Watch cites his work as “among the best out there”. With the launch of bellingcat Higgins is turning his hobby into a profession, successfully raising more than £50,000 on Kickstarter to fund its launch in August.
“It’s my attempt to solve a problem I’ve seen quite repeatedly in the last couple of years,” he told me. “I’m in an unusual position because I’m outside the normal circles of journalists and NGOs and activists, and they all come to me with these different ideas, and I find these great projects and tools and techniques that have been developed, but they’re just not getting them out there. They’d really work well together, but they don’t know they exist. There might be something an NGO is working on that a tech company could help with, but they’d never communicate with each other because they’re not in that circle. So what I’m trying to do is bring that stuff together in one site.”
In his work, Higgins uses a variety of technological tools to discover and assess evidence. The bellingcat homepage is a rolling almanac of global conflict, where “citizen investigative journalists” collect and post information as it arrives. Social media sites like Twitter and YouTube, mapping software like Google Maps, collaborative tools like chatroom/archive Checkdesk – these are all useful tools in sorting fact from fiction.
“That’s quite important because there are so many reports of things being moved from the site and messed around with. It’s useful to have these photographs recorded somewhere just in case there are discrepancies”, he says. “If we have these discussions on Twitter, in a couple of weeks’ time it’s impossible to find that original discussion, so having a place where it’s recorded and you can see the process of verification is useful. Also, I recognise there are a lot of people out there who aren’t experts, but keen on a subject and who can provide some good, useful contributions to investigations. I want to give those people somewhere they can go to and read about the tools and techniques they can use.”
The Buk launcher which is suspected to have been used to shoot down MH17 was photographed and recorded at different times on the day of the crash by local residents – Higgins reconstructed its movements, creating this map, showing it was within range of the plane at the necessary time.
It feels like there’s a paradox in our understanding of conflict, in the internet era: we’re seeing more of it than ever before, but the more we see, the less we understand. This is about more than Islamic State jihadis tweeting about Robin Williams one day and sharing videos of executions on YouTube the next – it’s about the idea of narrative in conflict, and our understanding of blame, and justice. We are bombarded with primary sources from war zones, but the traditional means of making sense of that information – things like war reporting by media organisation, or investigations by international bodies – are too slow, or missing altogether. It can feel like watching a documentary without a narrator. We know that what we see is meant to mean something, but what?
The crash of MH17 claimed 296 lives, and their fate was a miserable illustration of this. For more than 24 hours a grey train sat idling in the station of the small eastern Ukrainian town of Torez, waiting for permission from someone – it wasn’t clear who – to begin the journey of transporting the victims’ remains to Kharkiv, in government-controlled territory. We, elsewhere, could see the train. Wire agencies carried pictures of it, and those correspondents who were allowed near it noted that it was effectively unguarded. Nearby, the crash site was left open to journalists and rubberneckers to interfere with wreckage and the personal belongings of passengers alike – it was five days before the Dutch investigation team was allowed on site. The chaos of war and the breakdown of judicial authority was made immensely obvious, and it was heartbreaking.
It’s often said that the privilege of living in an always-connected world is that it has lead to the decentralisation of power. We’re no longer beholden to newspapers for news, record companies for music, Hollywood for movies, physical strength for ability, or governments for truth, and in its simplest form that’s a win for everyone who hates feeling like they’re being lied to by someone who knows more than them. Yet it’s damaging, too, to assume that all our new information sources are inherently better or worse than what came before just because they’re new – what we need, more than anything in this situation, is a decentralised way of dealing with a decentralised media environment.
With the crash of MH17, there was a surfeit of photos, videos and social media chatter directly from or about the crash, but very little in the way of clarity, little in the way of truth. What the conflicts of the current era show us is that developing new ways to handle the data we experience is as important as access to it.
“With MH17, the first photographs were taken two or three weeks ago, before the inspectors could get there,” says Higgins. “We’re documenting what’s there, and we can spot things which may have changed, which wouldn’t be obvious if you were an inspector arriving a few weeks later. For example, we’ve got five or six photographs of the missile launcher on the trailer, and we were able to find exactly where those were taken and what time they were taken and show that it travelled through 75km of rebel-held territory to the launch location, and then another 75km out towards the border, and that’s not something people on the ground could maybe piece together.
“[And] once we’ve established that, journalists on the ground could then follow that up. I think newspapers and news organisations need to be aware that there’s a lot of value in having this kind of intelligence work done for people on the ground, because then they can then be directed to more worthwhile stories.”
“She was kind of obsessed with a fan. It was initially on the ceiling, and she moved it, she moved it to the left, to the right. We didn’t understand why she was so obsessed with this fan. Until she recalled, when we did the walkthrough reconstruction, it was on the blades of the fan that she found human flesh.”
Israeli architect Eyal Weizman is showing me a video of a woman trying to remember the deaths of her family. She’s directing objects within a house – her home – in some off-the-shelf architectural 3D rendering software, the same software used to mock up the posters that hide empty construction pits from the sidewalk in cities around the world with images of laughing couples standing around island units in white marble kitchens. The aesthetic in this instance, however, is decidedly less upwardly-mobile – dirt floors, clay bricks and cheap furniture. There’s a fridge in the living room.
She lived with her family in Mir Ali, a town in Pakistan’s North Waziristan region, when the US drone attack took place in October 2010. Her husband and two-year-old son, and three other people, are now dead, and the building no longer exists. The trauma of the event has left her memory incomplete, and her ability to act as a witness has been compromised.
When the render is complete she “enters” it virtually, and her spatial memories revive other, hidden ones. The BBC’s Sherlock uses a stylised “memory palace” as a gimmick to explain Holmes’ exceptional powers of recall, but the idea has its roots in antiquity – classical orators could remember long speeches by imagining a walk through a vast mansion, with the objects in each wing representing specific memories. “When we are building the space where everything happened, in detail, every element – every piece of furniture, utensil, anything – in there, somehow you see how memory kind of re-emerges through these objects and returns through these objects,” Weizman explains.
A still from Forensic Architecture’s Mir Ali report.
This is the work of Forensic Architecture (FA), a research and activist group made up of artists, architects, cultural theorists and legal experts, all seeking to use the evidence of space and structure in the investigation of conflict. Weizman is its leader, having founded it with a grant from the European Research Council in 2011. He’s professor of spatial and visual cultures at Goldsmiths, and something of a notorious figure within his field. In 2002 he presented settlements as his choice of the best of Israeli architecture at the International Union of Architects Congress in Berlin, attracting the ire of the Israeli establishment. His work since then, in books like Hollow Land, has looked at the ways in which violence is conducted with the built environment, with particular attention paid to the West Bank and Gaza. He was a fitting choice of subject for Al Jazeera’s recent series of short films, Rebel Architecture.
The Architecture of Violence, dir: Ana Naomi de Sousa.
There’s a philosophical framework to Forensic Architecture’s work. It’s about “turning the gaze against the state”, using methods of spatial analysis as forensic tools in an “inversion” of the normal judicial process. He calls this forensis, not forensics. (It’s also the title of the Forensic Architecture project book, subtitle: “The Architecture of Public Truth”.)
“There is a feature that is a constant in working against the state with technology, and that is an inversion of the most important and basic forensic principles – that states, ie the police, should see in high resolution, see better than the criminal,” Weizman explained to me over coffee in his east London home. “Because otherwise you have a space of denial, or negation, in which the criminal can always mobilise. Now what happens in forensis, rather than forensics, is an inversion of the forensic gaze. It’s no longer the state and its institutions that are investigating citizens or non-citizens, but society, organisations, individuals and political groups. They are inverting the forensic gaze and looking up to the state, for state crimes. But you’ll always be in epistemological or visual inferiority. You’ll always have less of a resolution.”
He points to drone strikes as an example, where targets are chosen with cameras which have resolution of perhaps a few centimetres, or even millimetres, per pixel. By comparison, commercial satellite imagery has a metre (or maybe only half a metre) per pixel. For anyone trying to use satellites to keep tabs on war zones – be it an individual with Google Maps, or even the UN – the evidence is unavailable because of technological limitations. “What you see is that unlike cases in Darfur, or Gaza, where you can take before and after [images], the UN can take before and after images and see what has been destroyed, [but] in drone strikes the rocket goes through the roof, and leaves a hole in the roof which is smaller than the size of the pixel,” Weizman said.
“A before and after would not show you any difference. That is an example of an epistemological visual inferiority that you have to invert – you have to think, what access do we have? What information do we have? What new modes of thinking allows us to undo that space of denial between the few millimetre pixel and half-millimetre pixel? That’s the space of denial. This is why the state can say it cannot see it – they can say, we neither confirm or deny.”
Forensis is a method of forcing the state to admit to the existence of a crime, even if it won’t admit to culpability. Weizman points to the etymology of the word forensics, meaning “before the forum”, and how physical evidence demands a judicial response. If there’s no evidence, or the evidence is in an inaccessible form, then violence goes unchecked. “Our work happens in frontier zones. Forensis is forensics where there is no law, and in international law, the forum comes after the evidence, and evidence would call forth a forum. The graves of Srebrenica called forth the ICTFY [International Criminal Tribunal for the Former Yugoslavia] to gather around the evidence, for example. This is why the entire universe of international law is very, very fragile.
“It’s being formed around the type of violations that are happening, and it continuously changes as the nature of war is changing and the nature of violations change. New types of weapons, drones for example, or semi-robotic weapons. It requires a different way of evidence gathering, and a different way of presentation.”
FA’s case files show what this means. There’s the “left-to-die boat”, where 63 migrants died in adrift in the Mediterranean in 2011, within an area under the observation of Nato as part of the then-revolution in Libya. Satellite images of the area were used to establish boat movement patterns, and commercial boating data was used to remove the boats with known identities from the picture. Mobile phone signals were used to triangulate the migrant boat’s drift, establishing its presence near many other boats – and therefore providing evidence for the charge that the crews of those boats committed a crime by failing to aid a vessel in distress.
A reconstruction of the drift of the migrant boat, with other vessels charted in the area. Image: Forensic Architecture.
In another example Weizman shows me (as part of a work-in-progress, for a case being prepared for submission to the UN), a shaky handheld video leaked to MSNBC of the aftermath of a drone strike is broken down into a series of still frames. “We build a panorama,” he explains. “So now we have the entire ruin, we look at different features – we see a bend in the road here, we see a high tower here – which are then pieced together to create a full image.” The exterior features place the video in a specific place in Waziristan, while the shadows reveal the time of day the strike happened. The end result is a 3D render of the interior and exterior of the building, down to the locations of the shrapnel holes in the living room walls.
“And then you see here, two shadows where there are less shrapnel,” he points out. “We think that this is the shadow of the people, their body would have absorbed that shrapnel. They have been photographed onto the wall, no? Architecture and the dead body kind of combines here.” These kinds of sources, whether from traditional or social media, are enhanced and turned into evidence with technology – and, in turn, that evidence compels a national or supranational legal authority to listen to a case brought before it.
There’s a kind of cat and mouse game at play here, Weizman says, as every time the public learns to outpace the state, the state teaches itself new methods to reinstate the gap in resolution. “I think that [the state] shouldn’t have the upper hand, and I think social media mobilisation is something that could close that gap. But I think actually it’s never in the technology because the state could scan social media too. It’s in the aesthetic sensibility, or ingenuity in which you mobilise it. This is why our forensic agency is organised with artists, filmmakers and architects. There are no scientists – we are trying to think like artists, and trying to think about it with the aesthetic sensibility of art.”
The erosion of the advantage social media gives us is also a persistent worry with the work of Higgins and bellingcat, and why so much effort is spent on both verifying information and making the process of verification clear for anyone to check. There have been multiple reports of “cyber armies” funded by nations around the world – be they the hackers of the Syrian Electronic Army, the botnet legions of Russia, or Chinese Twitter accounts spreading anti-Free Tibet propaganda. (This is not something confined to geopolitical opponents of the West, it must be emphasised – intelligence agencies in the US and Europe are also “plugged in” to social media, sifting through it for relevant information, and it would be a surprise if they also were not running their own accounts to influence the online perception of breaking news.)
Higgins places his trust in the open source verification process as a way of eliminating the possibility of manipulated information, working in tandem with those who have proven themselves to be trustworthy gatekeepers. “These are individuals who build up a reputation, like myself, for being reliable,” he said. “People are collaboratively reviewing stuff, but you have the gatekeepers who do the final review, to make a judgement as to whether it’s reliable. If you’re a gatekeeper and you start putting out stuff that’s wrong, then you very rapidly lose your reputation. That’s the structure that’s developed, organically – the huge amount of information needs lots of people checking it, but also someone trusted to say it’s OK information. For example, if we’re looking at all this metal and the direction of the shrapnel, if we suddenly saw the shrapnel was coming in the other direction then we’d immediately be thinking ‘why is that different?’ That would be highlighted.”
He points to the last year’s work on the gas attacks in Damascus as an example of how “investigating all the things that it’s possible to investigate” creates a stronger case that’s more difficult to undermine with false information. “That involved tonnes of video footage, images, statements from people on the ground. When the White House published their map of who controlled the site on 21 August, the one area which was the likely launch site for the attacks was left blank – I was able to figure out, based on about 25 videos posted by a Russian-language news channel, where the government forces were. Then I was able to find footage from the opposition side attacking government forces in the area, which meant I could establish where the front lines were, and once I’d established the impact locations of the rockets that were fired we had a rough idea of what the range was, and we could say these rockets were in range of this government-controlled area.”
Another potential weakness in decentralised investigations is that their incompleteness can give a misleading impression – an example here being the attempts by users of reddit to identify the person responsible for bombing the Boston Marathon in 2013. It quickly became a witch hunt directed at brown-skinned men wearing backpacks in pictures of the crowd by the finishing line of the race, and fed into a larger panic over domestic terrorism which was eagerly encouraged by the tabloid press. One suspect in particular, a student named Sunil Tripathi, had been reported missing a month before the attack, and his family was distraught that their son was being named as a terrorist on the basis of one user mistakenly saying that they had heard his name mentioned on a police radio scanner. When Tripathi was found dead a week later, an apology from those who had jumped to conclusions did little to assuage their grief.
“It feels like there’s this frustration people have when they see this information on the internet and they don’t see it being reported in traditional news media, or if their government seems to be aware of it,” said Higgins. “It was interesting to see the response after the White House published their report on the 21 August gas attacks – the gap between what was coming out of the White House and what was coming out of my blog was absurd to a lot of people, especially when the UN report went on to agree with what I had.” He praises larger news organisations for “erring on the side of caution” when it comes to publishing source material that has been crowdsourced, and for limiting the loss of nuance as information gets passed from organisation to organisation.
The model for bellingcat isn’t just collaboration between individual armchair investigators, but also collaboration between traditional gatekeepers of information and those who are working outside of that framework – and Higgins believes traditional investigatory bodies will “really need to start engaging” with decentralised, crowdsourced methods of finding and analysing evidence.
Weizman, for his part, is keen to stress that architecture, like all technology, “can be used for good or bad things”, and finding ways to change the direction a piece of technology points is crucial part of his intellectual mission. Maybe – after all, the internet began as a military research project – we can see the work of crowdsourced conflict research as a kind of mob, storming the gates and turning something repressive into something liberating.
“The liberating is how you use it, and it usually happens at the moment of transformation,” he explains. “It happens in the first moment when you storm, or gain hold of, or enter a building that was used as a prison, a settlement, a military base and turn it into something else. I remember when the Palestinians first entered the military base in Beit Sahour [in the West Bank, where Weizman’s practice Decolonising Architecture has a studio] the first thing they wanted was to just smash it, there was something liberating about it. It was an incredibly powerful moment. The Palestinian police were trying to protect it and I thought it was wrong, because you need to let this spontaneous moment happen, that was the moment of transformation. And into those ruins we entered and we tried to convert it into something else.”