Forensic Architecture

Ann Mossop: Welcome to the UNSW Centre for Ideas podcast, a place to hear ideas from the world's leading thinkers and UNSW Sydney’s brightest minds. I'm Ann Mossop, director of the UNSW Centre for Ideas. The conversation you're about to hear, Forensic Architecture, brings together Michael Richardson from UNSW Sydney, with Eyal Weizman, director of the research agency, Forensic Architecture, at Goldsmiths, University of London. It was recorded live at UNSW Sydney. Enjoy the conversation.

Michael Richardson: I'm thrilled to be in conversation with Professor Eyal Weitzman, founder of the research agency, Forensic Architecture, and one of the most creative, determined and fiercely political activist scholars working today. Eyal’s work both through Forensic Architectures investigations, and in his scholarly writing has been a huge influence and inspiration for my own research. So it's a true pleasure to say welcome to the UNSW Centre for Ideas, Eyal Weitzman.

Eyal Weitzman: Thank you, Michael. It's absolutely a pleasure to be in conversation with you and your colleagues. And to reconnect after I think I haven't seen you for about two years, so, great to, you know, find an opportunity to speak together again.

Michael Richardson: Thank you. So in welcoming Eyal, I want to acknowledge that I'm speaking from the unseeded country of the Bidjigal and Gadigal peoples where I live and work, and where the grounds of UNSW Sydney are located, and also extend my respects to elders past and present and to First Nations people joining us online. These lands were stolen by the barrel of muskets and the violence of settlers, but also through the building of settlements, the fencing of lands for farming and industry, through the damming of rivers, the construction of infrastructure, and the imposition of colonial regimes of governance of registers and statistics, of the relocation of peoples from their lands to missions. This violence of settler colonialism was and continues to be inseparable from architecture and its techniques of surveying and mapping, of designing and modelling and planning. Eyal Weitzman’s work is about unveiling those forms of violence, and crucially about turning architecture’s techniques into subversive forensics for exposing and documenting state and corporate violence. Through his research agency, Forensic Architecture, founded in 2010, Eyal has investigated drone strikes in Afghanistan, dispossession of land in Palestine, boat turned backs of migrants in the Mediterranean, environmental racism and police killings at sites across the world and much more. He's the author of many books, including Hollow Land, the classic account of architectural and infrastructural violence in Israel, Palestine, Mengele’s Skull with Thomas Keenan, Forensic Architecture, The Least of All Possible Evils, and most recently Investigative Aesthetics with Matthew Fuller. 

So in this conversation, we want to get into some big questions about the nature of witnessing and evidence, and about the role of aesthetics in violence and investigation. To do that, we're going to hook our chat to some of Forensic Architecture’s cases, and show you some short clips as we go along. We'll do our best to narrate those so that anyone just listening can follow along too. I'd like to begin with some keywords that can maybe help us establish some common ground. And so those are evidence and testimony and witnessing, and maybe you'll have more. But to do this, let's watch our first clip from the Forensic Architecture archives. And in the interest of time, I will say that this clip, and the others we're going to look at tonight are truncated, and they race through the documentation of Forensic Architecture’s investigations fairly quickly. And so I really encourage all of you listening to go out to the Forensic Architecture website and look at the in-depth documentation there, to get a deeper sense of their work. 

So let's have a look at the first clip. This is one of Forensic Architecture's earliest investigations, called Drone Strike in Miranshah, which used shaky camera phone footage as the starting point for a granular reconstruction of a drone strike. So you'll see in the video here, and interest in what different types of sensing technologies, like satellites, can see and what they cannot see. Satellite resolution, for instance, can hide the human body. So what Forensic Architecture has done, is used this camera phone footage to go into some detail about what happens inside the building where the strike took place, to build a three dimensional model of the building itself and using shrapnel marks on the walls to identify the point of explosion of the missile, and also the places where bodies might have been located. As you can see here, the work is digital and rich in that way, but it is also material and designed to be presented in galleries and other such spaces. And we see too, the way Forensic Architecture is connecting this singular incident to a much wider form of violence in the use of drones in warfare. So, Eyal, perhaps reflecting on the clip we've just seen, and how do you understand these keywords, evidence, testimony, and witnessing?

Eyal Weizman: Yeah, thank you for this introduction and kind of leading us really into the heart of the problem. This is actually, I believe, the first or one of our first investigations, and through it really, we were able to set the terms for what there is to come. The investigation was actually commissioned initially by the UN Special Rapporteur for Human Rights. It was undertaken in, at a time, that the CIA or, the US administration was still neither confirming nor denying the presence of drone strikes in the Pakistan-Afghanistan frontier in Waziristan. Though images of destroyed buildings started to circulate. There was very little, it's nothing like, you know, in the early teens of the 21st century, it was nothing like the kind of media storm we are, social media storm we're in today, and particularly in that frontier region, where internet was not always connected. And particularly the siege over those areas, targeted smartphones and cameras, so people were going in and out of Waziristan were held at Pakistani checkpoints, and the media was quite literally stopped at the frontier of entering them. So that video, that on the basis of which the investigation is undertaken, needed to pass six hands, from one person to the other, to the other, to the other, until it arrived in Islamabad. And from there, it was broadcast by NBC, I believe. So it was an incredibly precious bit of evidence. And I'm saying that because when you receive a precious video, and precious not only because it's rare, but because it is filmed with a risk of life of the persons or the person that is filming it, and we can see that, we can see that the that video is dangerous to film by analysing the movements that are seen, the videographers own movement within that video.

And this brings us to your first question, testimony and evidence. Very simply stated in legal context, evidence refers to the account of objects. Testimony refers to humans. There is a kind of a human non human divided across the evidence testimony definition. Though, with videos such as this one, evidence and testimony get entangled. Videos record from both their ends. When something is in front of the camera eye, the lens, you know, the videos, you've seen, records of ruin, it records the shrapnel in a room. But it also records every video records the movement of the videographer, the position and movement of the videographer taking those. It's almost like if, if you like, what is being recorded in front of the lens is the evidence, and the testimony i.e the, kind of, the condition of the subject recording is what is implicit in each one of those. But it's almost like a video is a kind of a halfway mirror in which the subject and object gets superimposed in a particular way. And what is absolutely essential and where we have recognised that that video was filled with a danger of life, that it's a kind of, parrhesia, you know, in a sort of sense that Michel Foucault gives it, a kind of a testimony that requires risking one's life to deliver, is when we see the person recording out of the window. One of the sequences is the videographer records out of the window, but we don't only see the landscape of destruction outside of the window, we see the window frame in the picture frame. And when you think of it architecturally, you can locate a videographer deep within the room, filming from the shadow of the room, outside, that videographer does not want to be seen videoing the strike. We don't know what, precisely, is the videographer wary of, whether it's a double tap, which is a very common strategy in drone strikes, by which you fire one missile and you wait for people to gather, and you fire the other one. Or it is, perhaps, that people on the ground, the people being targeted, the people feeling vulnerable and threatened by that would suspect that videographer to be spying on them. But  that is an incredibly important moment. The image frame contains the window frame, a framing side of frame is a testimony to the condition of filming itself. And we see it again and again and again that filming is not simply being a witness. It's not being a citizen journalist, etc. It's risking one's life. In Palestine, an analogous to that is the fact that Palestinian filming Israeli incursions into Gaza, are risking their lives to do it because the shoot to kill instructions to Israeli soldiers during the Gaza attacks, are to effectively shoot at anyone directing a camera at them. So we understand also, that cameras are part of the conflict. They're not simply a question of representing what has happened, they are part of the event as it unfolds. There is an event itself, the event or photography of videography itself, is incredibly important. 

So this is one. Second is what you alluded to, the threshold of detectability. When buildings are struck normally what investigators would do is a kind of before and after, comparing before and after photographs from satellite images. There is a certain kind of rhythm by which satellites pass over a place, creating the bookend and event. You do not capture… the satellites never, rarely, capture the event as it unfolds. But the event turns into an architectural event by comparing the state of a building before and the state of a building later. Built, destroyed, we can see in those kinds of montages, these sort of like, dialectic montages, to borrow a term from Eisenstein. You could see the, sort of like, the politics turns into architecture in that way. But also environmental conditions, icebergs melting, forest, destroyed, floods, etc, etc. We know that. The problem with drone strikes is that the smallness of the munition, the hellfire missile that is fired by drones, is smaller than the single pixel in which the satellite image captures the ground. Every pixel is about half a metre by half a metre. Not because this is the actual optical resolution of satellites, we know since Donald Trump stupidly, and even in violation of US classification secrecy, tweeted a satellite image, I think back in 2019, of a site in Iran, that the resolution is much sharper than half a metre by half a metre. Today, 30 centimetres by 30 centimetres. It's much sharper than that. The resolution stops at that because half a metre by half a metre is a, kind of, a modular, it's designed according to the human body. Human body from above is about half a metre by half a metre, and the pixelation is meant to erase people out of the image. So the crime, that is to say, the crime of targeted assassination, in which buildings are targeted by missiles smaller than that, falls under the threshold detectability of the means by which we can see them. You cannot do satellite image analysis on drone strikes. You need to resort to other things. And then you get those, what we call, weak signals, blurry images, delivered under the risk of life from those environments. And you need to, because these are weak signals because we encounter forensics always work with lesser optics, lesser technology than the technology of the perpetrators, right? The US sees in high resolution, their optics, their missiles, their entire, kind of, analytics have got much more data than us, we work with very few and weak signals, we need to creatively work with them in order to see things that perhaps others could not. And here is where imagination is playing a role. And in that case, we have understood that we could take the walls of the room and I think this is what you've shown the interior of the room where people were killed by the drone strike. Think of it as the architectural interior, as if it's a photograph. Exposed to the blast, in a same way, or in an analogous way in which a negative is exposed to light. The shrapnel on the wall had shadows. And those shadows are where people stood. So effectively, a kind of a human photograph, think of it like, you know, in an analogous way, also to the way in which the photographs of people in a nuclear strike on Hiroshima, no? Would kind of like, imprint on the staircases. The violence of the blast rendered the human figure on the architecture. Very powerful moment for us in understanding the kind of augmentation of different, sometimes banal material surfaces, such as walls, such as leaves, such as the sea. We will talk about all of them later, as hypostatised mediums that register continuously if you know how to read. And if you care, to look at those very, very studiously, very carefully, very patiently, in order to extract the human from the rubble.

Michael Richardson: I was struck as you were talking about this idea of the threshold of detectability, which you've said in your writing hovers between being identifiable and not where things sit in this in between space. And I'm interested in the way through Forensic Architecture’s work, the thresholds of detectability, and the thresholds you've pursued have moved, maybe from more built architectural spaces, like we just saw, you know, a home that had been destroyed by this terrible drone strike, right? So sort of a recognisably architectural space in the, kind of, way most of us think about what architecture is, or what buildings are. But Forensic Architecture’s work has extended far beyond built spaces like that. And so I wanted to show another clip now of an investigation you did on herbicidal warfare in Gaza. And so again, this clip starts with – investigation, I suppose – starts with video footage carefully taken at risk by aid workers, I think in Gaza, who had seen a plane dumping some kind of herbicide or some kind of liquid out of the back, as it flew through the border area. But the analysis that we're seeing snippets of on the screen now, it shares some resemblance, of course, you're using satellite imagery and other types of aerial views. But we're in a space of meteorology, of weather, of the movement of wind, of the biological material that is damaged. So I wondered if you might say a little bit about how the threshold of detectability has moved and transformed for you and for Forensic Architecture.

Eyal Weizman: I think that you know, the idea of operating under the threshold of detectability is really structural to our work. And in order to explain why it's structural we need to speak about what counter forensics is. So initially, Forensic Architecture is, we don't see ourselves as a forensic agency, but a counter forensic agency. And counter forensics is not simply taking the means and technologies of forensics as it has developed as a police practice since the late 18th and 19th century, but belonging and requiring another type of work. Initially counter forensics is always investigating the investigators, only investigating state violence. We always work directly with communities, such as the communities of farmers that you have shown now in the Gaza border. The area along the border in Gaza is one of the only free open spaces left for cultivation in an otherwise extremely dense or, you know, forcibly dense, because people are not allowed to move out, very, very, very difficult to migrate out. Obviously, the, kind of, the Gaza Strip is incarcerated under a very punishing siege, they have to rely on cultivating sustenance, field and cultivation there in order to be able to resist the very frequent stops in the kind of humanitarian trucks that Israel would agree to allow in. So we are talking about a very fragile agricultural frontier that is cultivated along the border. Working together with these farmers that are effectively suffering, a kind of invisible form of violence. 

What has occurred there is the Israeli military interest is what they call to expose that frontier area. So from their perspective, they have destroyed buildings along that border, so that the resistance would not have any way to hide as they approach or even protesters would not have any kind of refuge and would always be exposed to Israeli snipers, on the other side, that area was initially cultivated with a citrus, Gaza was very famous for it’s citrus. Those have been uprooted. The greenhouses are routinely being destroyed, and therefore Palestinians have to cultivate on the surface. Our Palestine researcher, Sridhar Mulleavy has actually worked for years with those farmers on trying to understand how we can visualise that invisible violence of mobilising the air, mobilising toxic clouds, to destroy even that very, very low scale, a kind of agriculture that is happening along the border, we have placed different kinds of sensors on the ground together with them. It's almost the analysis, the kind of the sensing of the thing has been done with the Palestinian farmers along the borders, who have recorded the wind, who have recorded levels of toxicity registers on the leaves themselves. So here, we have different kinds of leafy vegetables that are cultivated along the border that show the imprint of toxicity on them. So now the surface of the leaf becomes a photographic surface that registers that violence. But we needed to connect that hard evidence of contaminated leaves with an Israeli activity. Now Israel was spraying its herbicide on the Israeli side of the walls and fences around Gaza. But it was doing that at a time, everyone that lived or visited the Mediterranean region knows that in the morning, wind goes from the land to the sea, ie from Israeli controlled area into Gaza, and then in the evening comes the breeze from the sea to the land. So if you actually spray herbicide in the morning, you weaponize the wind to carry toxicity from the Israeli into the Palestinian control sides of the Gaza perimeter. 

And here comes another attempt or another kind of line of thinking that we've developed around airborne violence. It's connected to our work on tear gas to our work on different kinds of chemical strikes in Syria. It's connected to our work on forest fires, and we even looked at the forest fires in Australia as part of that. Attempting to understand the clouds, sometimes visible, sometimes invisible, as a different kind of object than the hard objects that forensics likes very much to deal with. You know, one of the principles of forensics is that every contact leaves a trace, it's called a Locard principle. And indeed, that is right with hard surfaces. Hard surface hits a hard surface, some trace is left, even if it is hard to read. Clouds are kind of a limit concept for forensics, they are the epitome of transformation, of metamorphosis, they are moving. In order to understand their form, architecturally, as a dynamic form, you need a combination of mathematics, mathematical simulation, evidence, ground truth, where you have the evidence of what material it is, what's registered and leaves, we see where it comes from, and then we need to understand the morphing architecture of those clouds. We call that work cloud studies. It is a long term project that we invest enormous amount of attention in order to map that airborne violence. And we've understood that when you think about clouds, whether they are tear gas, chemical strikes, those kinds of herbicides that are moving with the air, there's two ways to think about it. One from the inside, you're inside the fog, a cloud is an optical condition of visibility, or fractured visibility, of obscured visibility, but also a kind of an epistemological space, of blur, of not knowing where things are. And another way is looking at clouds from the outside as constantly morphing objects. And funnily enough, we, you know, we developed our technique by looking at the way in which the history of art actually dealt with and attempted to, kind of, capture clouds as, as these kinds of continuously metamorphosing objects faster than the hand of the painter can actually capture them. It's another way in which counter forensics needs to, actually kind of like, draw on resources that are not only scientific, but coming from the cultural domain, such as art history.

Michael Richardson: I was thinking as you were speaking of nuclear clouds, and other ways that air and atmosphere have been weaponized. And I was thinking too of, nuclear testing here in Australia, which was conducted in the 1950s and 60s primarily in Maralinga in South Australia, where nuclear blasts transformed the sand into, like, hardened silicate, like, almost glass, but where a lot of the damage was from the nuclear material, the irradiated air, as it spread across what traditional Aboriginal lands and damaged plants and animals and people and so on. And yet, so much of that is difficult to detect. I'm wondering, with your pursuit of the thresholds of detectability in different sites and spaces, an audience member submitted what I think is a really interesting question, which is, do you feel like, as your work pursues those different thresholds and develops these techniques for revealing violence below this threshold of detectability, or at the threshold, is there the possibility that state actors and other people learn from those techniques and, like, develop their own counters to your counter forensics? And so is there a, kind of, risk that there becomes a struggle over misinformation or over or over truth in the work that you do that extends beyond like the specific case and it's about the process itself?

Eyal Weizman: No, absolutely. This is, this is something we encounter all the time. Initially, we had different police forces, police watchdogs, military, even, the British military, the British Ministry of Defence asking us to undertake work for them and we always flatly refused. We do not work for states. We never knowingly took a commission from a state. But the public domain is large and chaotic and information that is out there could be digested and interpreted and learned from by actors that you want to have nothing to do with. As technology evolves, also techniques of analysis evolve. I think that for us, we need to first acknowledge that as a structural problem, because we would put all our investigations and our technology, as open source, that they could be used. It puts on you a certain kind of demand that is very extreme. And that is to say, we need to continuously develop our techniques so that at any given moment, to our estimation, we are able to do more with less evidence than state agencies. So we see that, whenever we have the opportunity, many times they try to disqualify it from coming into court for coming into debate with state agents. But whenever it comes to a factual analysis, whenever it's come to uncovering a case, on a factual level, we prevail, I mean, if they want to, I mean, they manage to succeed simply by disqualifying us calling us activists calling us artists, calling us, I don’t know, all sorts of other, kind of things. And the reason is that, each one of the cases that we take, we take not because we know how to do it, not because we will apply the same technique that we have applied before. But we take projects only in as much as we can develop new techniques from it. Only in as much as we could evolve, we never take projects we know how to do. So it requires you to constantly innovate in order to momentarily be above what you think police force here, the Metropolitan Police in London or the Israeli army can do with the evidence, and therefore we, you know, we can catch them when they lie, we can make claims that actually contradict or expose their violent negations. Because we know that negation is not simply a rhetorical layer that is put on violence. It's not an add on to violence, negation enables violence. The negation from when the Israeli police or military negate killing of a young Palestinian boy, they say that did not happen. And the negation of the Nakba that started in 1948, are connected to each other, you cannot maintain a regime of violence if you do not negate its effect continuously. So violence and its negation are entangled and therefore the continuous confrontation with and working on the, kind of, the material texture of the history that we have, and we see those videos and satellite images, leaves walls, as historical documents in telling history is important in, kind of, confronting regime of violence and domination.

Michael Richardson: I wanted to ask a little bit about your latest book, which is called Investigative Aesthetics: Conflicts and Commons in the Politics of Truth, which you co-authored with the media theorist, Matthew Fuller, media theorists and many other things, Matthew Fuller. It's a terrific read. And I'd like to dig into it a little bit. And I wanted to start with this crucial word, aesthetics, which you've mentioned several times in the talk today. And when you say aesthetics, you don't mean the apprehension of beauty. You don't mean aesthetics, in the simple thing of something that is pleasing to the eye. So what do you mean, when you say aesthetics?

Eyal Weizman: I guess, you know, Matthew and I, and others in Forensic Architecture, when we say aesthetics, we mean close to what the people that use this word in ancient Greece meant, it's that which lends itself to the senses that which gets registered, in a particular way. So they kind of like the base level of aesthetics for us, is the idea that surfaces, the surface of the table next to you, where you sit, the chair, even the surface of your screen that you're looking at, is continuously registering its environment. It's perhaps a trace that we cannot see with the naked eye. Perhaps the kind of analysis needs to be much more careful. Perhaps it involves heat sensing on the table, to see how the table actually registers your heat, or how a leaf registers pollution, or how a surface of the building registers proximity of humans within it. Or as we've seen before, a perfect example of aesthetics, the interior walls of the room in Miranshah that was struck by a drone became aestheticised to the blast. I.e, the blast got registered in it. So that's the base level of aesthetics. 

But of course, aesthetics is important for us in Forensic Architecture, because it's the means by which we start composing. And every case is a composition of relation between different aestheticised surfaces, the leaf that is aestheticised to the toxic cloud, the concrete room, or the concrete wall that is aestheticised to the blast, the sensor on the smartphone that is aestheticised to light, a satellite image, the ground itself, the air, all of these are kind of aesthetic media. And the case is about composing a certain relation between them and making that composition. One that matters, one that is political, one that has a claim to make. Otherwise, it’s an infinite set of relations between things. Combining one to the other, this photograph, with this video, with this satellite image, with this leaf, with that cloud, together with this memory of an event. Think about violence, in the sense that, you know, an army grades village, a town, think about, you know, Ukraine now, every surface registers, people remember in fragments, traumatised sometimes by the extremity of violence, the ground register, the tanks, the chains of the tank, the air register, the suit, or the chemicals that are being used as part of specific munitions, etc, you need to compose them, that act of composition is done sometimes on an editing suite, sometimes on an architectural model, is an aesthetic composition. And then the last layer of aesthetics is when you take that composition into a forum. And you need to make claims with it. And you know, of course, every forensic specialist knows how important aesthetics is to make a good argument, just like every lawyer understands that the force of rhetoric is an extra beyond the words that are being used as a performativity in the aesthetic to a presentation, because what you need to arrive at is conviction in both senses, both to convince somebody and to convict somebody, if this is where you operate. And then you know, Matt and I evolve it into many other levels. We have a whole, kind of, travel through the notion of aesthetics through categories, such as hyper aesthetics, which is the way in which you augment the aesthetic surfaces of registration, and then connect them together. But also the failure of aesthetics with hypoesthesia. When registration no longer makes sense, where you get so much information that creates short circuits to new understanding, a certain aesthetic of madness, of collapse, in which the amount of information starts erasing itself.

Michael Richardson: I'd love to get into both of those ideas, perhaps through the example of the Ukraine. But before we get there, I do want to invite you, I'm going to show a clip from an investigation called the destruction and return at Al-Araqeeb, in the Negev region of Palestine. So if we could play that clip, I was struck in exploring this investigation, and what you're seeing here is a very small snippet of a large amount of very powerful documentation, over many years of working with the community there to identify different aspects of the history of living in that land. And I was thinking about this arc of history in your work in thinking about colonial violence here in Australia, too. And the contestations that we have had over the history of colonial violence and of Aboriginal inhabitation of the land long before white folks like me and my ancestors arrived. So I just wanted to invite you really to tell us a bit about the importance of Palestine to your work, and perhaps your thoughts on this longer arc of history and the potential of Forensic Architecture to explore it.

Eyal Weizman: Thank you for giving me this opportunity. I think I think you all know that I’m Israeli born, and somehow the, kind of, the origins of Forensic Architecture emerged in the anti-colonial struggle against Israeli settler colonialism across Palestine. In one of those sites is the desert, particularly both symbolically, but actually, physically, one of the most violent frontiers in Palestine is not really organised around the walls in Gaza, or the walls that is being built through the West Bank on Palestinian land, but a certain invisible border between areas that are considered desert and area that are considered sown. And that line that defines that border is a certain meteorological line, 200 millimetre rain per year line, south of which according to Israeli interpretation of environmental data, you will not be able really to cultivate wheat without artificial irrigation. That is, anyone that lives in the desert, everyone that lives under the 200 millimetre rain per year, does not own the land in which they live. It's very much like the kind of Terra Nullius that was a prominent feature in Australian colonialism. We call it the dead land doctrine, in which the desert belongs to no one hence, it belongs to the state. However, the desert is, or that which is called a desert, even the word desert is politicised, in that sense, that region of the Naqab is inhabited by those survivors of the displacements of the 1940s, where the majority of the Palestinian Bedouins from the Naqab were expelled into Jordan, Egypt. Those that remained hold on to that very fragile frontier and face continuous harassment and continuous settler colonial violence in being expelled and Jewish settlements being built, military bases being built, polluting industry being built on their lands, and even a nuclear reactor is being built in that part. So that frontier is about thinking about the role of the environment in the, kind of, the history of the Nakba. thinking about desertification, in our part coming from the south to the north, as environmental pressure increases, thinking about settler colonialism trying to push the desert southwards through that category of making the desert blue, and kind of, you know, almost a messianic kind of imaginary that is not only Israel, it's very much associated with Zionism, but was also the policy of the French and Italians as they were taking over, in the beginning of the 20th century, North Africa, trying to make the desert blue. Between those two forces, settler colonialism and climate change, lie the Palestinian Bedouin population in this part, that worked together with us, and here is a kind of a very interesting sort of methodological point to make. When you undertake counter forensics, you try not to put yourself as an external neutral expert, you try to create a certain engaged objectivity through the category of what we call open verification, which means, big collaborative networks that are led by the people at the forefront of struggle. In this case, particular Bedouin families of the Al-Touri and Al-Ukbi, resisting those two frontiers, desertification, settler colonialism pushing from two sides. The attempt to erase their history through Israelis planting trees on their old lands, erasing evidence, those Bedouin Palestinians are expelled and return into those lands, actually, in the hooks of returns are the archaeological evidence of the previous presence there, including cemeteries, including dams, including stone houses, etc, that they had those needed to be mapped and put together with oral histories, to create an understanding of the actually inhabited desert, a desert that is, that is actually full of life and full of different types of occupation, and different forms of life that are kind of lost now, under settler colonialism. [To] bring it out from under the surface required that deep collaboration between meteorologists between us, between the, obviously, the villagers themselves, in creating that evidence that you've seen in this clip.

Michael Richardson: Turning now a little to talk about the Ukraine, the invasion by Russia more recently, but also, I'm interested in this in, obviously, it is a horrific human tragedy. But one of the things that we've seen during this conflict is the huge amount of visual material that we’re encountering. You know, not just smartphone imagery of violence and of damage to buildings and homes and bodies and so on, but also drone footage, imagery through the sites of weapons targeting systems, satellite imagery, there's just a, kind of, overload of this type of imagery. And I thought that, it seems to me, goes hand in hand with this shift that's happened in the last 10 years, I guess, over the life of Forensic Architecture, in some ways, the emergence of machine learning, the rise of big data. And so these like, new techniques of analysis, which can have their problems in many contexts, but Forensic Architecture has started to use in some really powerful ways. So I thought we might watch a clip of a tank investigation that you guys did in 2019. This was commissioned by the European Human Rights Advocacy Centre to verify the presence of Russian tanks in eastern Ukraine in 2014, at a time when Russia had claimed that it was not directly supporting separatists. And so the investigation if anyone wants to follow it up, it's called the Battle of Ilovaisk. And I was thinking about this investigation in which you use machine learning analysis to deal with a huge amount of video footage, and other types of imagery from the conflict. But you also, as we're seeing on the screen here, built these 3D models of tanks, and use those to help in your analysis. I was curious about your use of machine learning and these types of techniques. And also the way you've made those into platforms like we're seeing on the screen now, that people can access. And I was wondering about that in the context of Ukraine today. And I know, you've got a real interest in what's happening there. And so I wondered if you might just talk a little bit about the use of these new kinds of techniques about how to deal with a conflict of this speed and scale that's happening in real time, wherever you want to take that Eyal.

Eyal Weizman: No, it's a great opportunity to speak about, effectively, what I'm going to do five minutes after we are done with his talk. I mean, we are very worried and working very hard on Ukraine, particularly through collaboration with a very wonderful Ukrainian group of investigators called The Centre for Spatial Technologies. And, you know, speaking about hyperesthesia, hyperesthesia before, the amount of information that comes, I mean, you're absolutely right, is beyond the capacity of any human researchers to make, now, sense of. And really, it is about the shifting condition of our analysis. If, if, kind of like open source investigation is very much associated with the early years of the Syrian civil war, or the repression by Assad regime, you know, there were one, two, three, slowly, there were, kind of, four or five videos around incidents of interest. Today, we have hundreds, if not thousands, of videos around any incident that needs to be verified and analysed. And they're no longer very short, they're no longer 10 seconds long, they are hours long. They are, you know, the streaming videos, sometimes three, four hours long. Our researchers are unable to actually just the time of watching it, there are many more hours of the Ukraine conflict, then there are hours of conflict… there are many more hours filmed, sorry, of the Ukraine conflict. So what we need to undertake is employ machine learning on the first layer of analysis, and we have developed what we call synthetic classifiers, that is using our capacity as modellers, to actually train algorithms to see particular objects of interest. So let's say you want to identify a particular type of tank or particular kinds of munition, typical type of plane, particular kind of object, any object, you need to show it to the algorithm, quote, unquote, show, 10,000 20,000 times from different perspectives, in different lights, in different resolutions, until they kind of, classify, get it. It's almost analogous to, kind of, teaching a child to see, and to understand the concept of an object beyond the specificity of a particular object. We use our ability as architects to model and to render realistic images, to almost trick our classifiers to say, oh, these are real images, and we can generate those variations very, very fast. And then we send those classifiers out, what we call, on the wild. They scan the internet, and they come back with hopefully, all points within you know, thousands, tens of thousands of videos where a particular tank or particular munition is being seen. Still, most of the things that come back are false positives. But there is hope, we see this thing is working. 

What is interesting, and really where I want to kind of like, a question that I want to engage with, towards the end of this lecture is a critical introspection of technology. Of course, we use, as you saw at the beginning, we use satellite imagery, we try when we use such imagery, which has origins in the Cold War, in military history, etc. We try to do two things, we use it to expose those things that we want to expose, to undertake our investigation. And we use our investigation in order to expose the biases and problems and politics within those technologies themselves. Each one of our investigations, we hope it to have a kind of a Janus phase. We’re looking at the technology, and we’re looking… we're looking critically at a technology, a violation of human rights that's happened in this case, within machine learning algorithms, and then we employ machine learning algorithms to expose human rights violations in a real space. How is that done? A lot of the problems that we identify with machine learning itself is that they're kind of black boxes. We don't know how they work, it's very hard to reconstruct the internal complex computational process that happened within machine learning. But when you have synthetic classifiers when you control the input, and see what comes in the output, obviously, it allows you to identify Russian tanks in Ukraine. But it also helps reflect on how the algorithm is built, what biases exist within it. And we've discovered all sorts of interesting things about those machine learning algorithms simply for critical use. We realised that it's not really about modelling realistic Russian tanks and showing the computer to do it. But kind of like what we call extreme images. If we take a model, a 3D model for Russian tank, we'll paint it in zebra stripes, pink and yellow, and render it from all sides in different kinds of strange lights, those extreme outlines of images would help identify a, kind of, more like, the sort of, like mainstream and more common images that that could be expected to see, something that is counterintuitive. But also explain why social media companies require extreme behaviour online in order to train predictive algorithms on people. That the extreme behaviour actually is what is necessary for those algorithms to do it. So. So here, you have two kinds of human rights violation, one happening on the ground by Russian tanks, entering, murdering, destroying cities and villages in Ukraine, on the other hand, our work allows to, kind of, expose the biases and potential human rights violation within the algorithms themselves.

Michael Richardson: Fascinating. One more quick question, which to me is one of the things that follows from what you've just been saying. But also in reference to your work with partners in the Ukraine, through your practices of open verification, and your willingness to put the tools that you develop up online. And that is to ask you, what you mean by an Investigative Commons? And what might it mean for all of us, for people to participate in an Investigative Commons? Is there a call to arms or a call to Commons perhaps, that we should be heading?

Eyal Weizman: I think every investigation is also an evidence to the social relations that made it possible. And we need to stop thinking in terms of the, sort of, neutral objective expert look, with a view from nowhere on to an incident. And think about embodied form of knowledge, situated intelligence, and working together with the people that are leading the struggle, with the people at the forefront of any struggle, whether they're resisting Israeli settler colonialism, whether they're resisting police violence in inner cities, whether they're resisting the Russian military in Ukraine, though, of course, not the Ukrainian military, we will not take a military side to that will always worked with civil society in a kind of very diffused way. So an investigative common is really the relation between the people on the ground, activists next to them, maybe the lawyers, dispersed and wide network of experts, perhaps even because, you know, we may show the work in court. So the lawyers, or we show it in a gallery, therefore, the curators, the publishers, the distributors, create a certain kind of momentary network, a community of practice, a, kind of, a social contract emerges around each one of those investigations. And I think that this is what makes them political. What makes an investigation political, is not only that you investigate political processes, it is not only that your finds can change, maybe political outlook on things or intervene in political movements, etc. But that you create a certain Commons, a certain shared community of practitioners that work together in order to describe that meta political condition, which is, you know, the facts that surround us, the real, what we call. That network of practitioners, the idea that verification we don't like the word truth, because it's very mono perspectival, and rather than imminent, it's transcendent. You know verification is turning truth into a verb. Verification and open verification is precisely breaking that idea of expertise and working in big collaborative networks. And sometimes paying the price for that when we’re called into court and they say to us, your evidence is contaminated. We say that is the way to produce evidence today, through relations. It requires a wide network, it requires the situated knowledge of people on the ground. So I would encourage anyone that wants to use techniques such as ours and other open source methods online, and understanding that the ethos of those works is embeddedness, situatedness, and wide network of participants.

Michael Richardson: Well, thank you Eyal, for your time today, and for that call. I encourage anyone who's listening and has not already done so, to get your hands on a copy of Investigative Aesthetics: Conflicts and Commons in the Politics of Truth, that Eyal wrote with Matthew Fuller. It's from Verso Books, I encourage you to support your local bookseller, buy from Verso direct, or via the link on the event page for our conversation. It's a really wonderful read and suffused with creativity, insight and imagination that we've heard from Eyal today. So thank you Eyal Weizman for your time.

Eyal Weizman: It was wonderful speaking.

Ann Mossop: Thanks for listening. For more information visit centreforideas.com, and don't forget to subscribe wherever you get your podcasts.