1 00:00:00,000 --> 00:00:15,015 *34c3 intro* 2 00:00:15,015 --> 00:00:19,450 Herald: Okay, so again, once again: Good morning to everyone! We are really sorry 3 00:00:19,450 --> 00:00:23,940 for the delay, but as you know this is the Chaos Communication Congress. Chaos is 4 00:00:23,940 --> 00:00:28,930 part of the show. In any case, we are ready to start our first speaker for the 5 00:00:28,930 --> 00:00:33,270 day. This is a talk that I'm personally very excited about. Very interesting in my 6 00:00:33,270 --> 00:00:38,510 opinion. Our first speaker is an architect. He's a professor of spatial and 7 00:00:38,510 --> 00:00:43,649 visual cultures, and the director of a forensic architecture at Goldsmiths 8 00:00:43,649 --> 00:00:48,530 University in London. Please give a big round of applause to Eyal Weizman. 9 00:00:48,530 --> 00:00:51,140 *applause* 10 00:00:51,140 --> 00:00:58,099 *microphone off, inaudible* 11 00:00:58,099 --> 00:01:03,699 Eyal Weizman: excited to be here. My first congress, atually, and I'm sure it's gonna 12 00:01:03,699 --> 00:01:10,030 bring good luck. The fact that we had some technical equipment problem right now. I'm 13 00:01:10,030 --> 00:01:15,851 running in London something that is better described as a counter forensic agency, as 14 00:01:15,851 --> 00:01:20,150 a civil society counter forensic agency. There's no better way to explain what 15 00:01:20,150 --> 00:01:27,109 counter forensics is, a certain turning around, repurposing of the forensic gaze 16 00:01:27,109 --> 00:01:34,040 towards the state, then looking at a series of issues where security forces or 17 00:01:34,040 --> 00:01:39,530 the police are the perpetrator. So, what I'm going to show you today, very fast, 18 00:01:39,530 --> 00:01:44,579 are three cases. The first one in Israel, second one in Germany, third one in Mexico. 19 00:01:44,579 --> 00:01:51,610 Each one involves violence or alleged violence by the police and each one also 20 00:01:51,610 --> 00:01:59,100 involves a different mode of research and technique, in doing so. First one is a 21 00:01:59,100 --> 00:02:07,310 place where I come from, Israel Palestine, and the issue is really the force eviction 22 00:02:07,310 --> 00:02:13,920 of Bedouin communities that have been living in the north part of the Naqab or 23 00:02:13,920 --> 00:02:21,540 Negev desert in the south of Israel, for generations now declared by the state to 24 00:02:21,540 --> 00:02:28,770 be illegally occupying those places and are subject to continuous raids by which the 25 00:02:28,770 --> 00:02:32,570 police... Oops, you don't see the slides. 26 00:02:32,570 --> 00:02:40,070 Slides? Slide? Guys? Okay. Sorry, one second. 27 00:02:40,070 --> 00:02:43,050 *Inaudible* 28 00:02:43,050 --> 00:02:45,450 *laughter* 29 00:02:45,450 --> 00:02:48,181 All right, so I'll tell you a little bit more about forensic 30 00:02:48,181 --> 00:02:54,930 architecture until until they do so. Basically what we are is a group of 31 00:02:54,930 --> 00:03:03,870 architects, filmmakers, some investigative journalists, coders and we join together 32 00:03:03,870 --> 00:03:10,830 to create a forensic agency. It was a kind of an experiment that we 33 00:03:10,830 --> 00:03:23,770 started around 2010, because we felt that both,technological and political changes 34 00:03:23,770 --> 00:03:30,530 enabled and demanded a form of counter forensics that we are now practicing. 35 00:03:30,530 --> 00:03:39,230 Initially, we were working in conflict zones. Towards, or starting, really, in 36 00:03:39,230 --> 00:03:45,820 this millennium, in the year 2000, war became an urban phenomena. Almost all 37 00:03:45,820 --> 00:03:52,140 conflicts that we were looking at took place in cities, in and amongst buildings. 38 00:03:52,140 --> 00:03:58,930 So in a very straightforward way, just like a medical doctor, a physician can 39 00:03:58,930 --> 00:04:04,800 turn into a pathologist, we architects, we're turning the tools of our trade, 40 00:04:04,800 --> 00:04:09,220 understanding of building, a way of interpreting its materiality, its 41 00:04:09,220 --> 00:04:15,380 physicality into an evidentiary technique. So initially, what we were doing was a 42 00:04:15,380 --> 00:04:22,200 kind of archaeology. Archaeology of the very recent past, or archaeology of the 43 00:04:22,200 --> 00:04:31,750 present - this is a term by Gilles Deleuze - and looking at piles of ruin, 44 00:04:31,750 --> 00:04:37,080 of a building that was destroyed in a bomb or in a firefight, and trying to read from 45 00:04:37,080 --> 00:04:47,060 the disposition of the rubble something about what has taken place within 46 00:04:47,060 --> 00:04:55,900 and around it. So it started as a kind of an archaeological practice. But then 47 00:04:55,900 --> 00:05:02,290 again, very often, when we start working in conflict zones, and the first projects we 48 00:05:02,290 --> 00:05:11,650 had were in Palestine and then in... we were working to uncover drone warfare in 49 00:05:11,650 --> 00:05:16,530 the Pakistan-Afghanistan frontier, we realized very fast... 50 00:05:17,530 --> 00:05:20,579 Technician: Sorry, could you... *inaudible* Eyal: Should I what? 51 00:05:20,579 --> 00:05:24,160 Technician: The presentation on that stick so you can use this laptop. 52 00:05:24,160 --> 00:05:27,190 E: Oh, this gonna take 20 minutes to pass them. 53 00:05:27,190 --> 00:05:29,690 T: Really? E: Yeah. 54 00:05:29,690 --> 00:05:37,840 I mean, because I started to put it on a stick, and... Listen, can you connect this 55 00:05:37,840 --> 00:05:41,420 I'm gonna do something else... Technician: This is connected. Can you... 56 00:05:41,420 --> 00:05:46,600 E: If you can get an image from this, I'm just gonna run things from my website. 57 00:05:51,090 --> 00:05:57,850 So, as I said, the first level of our work was archaeological. Material archaeology, 58 00:05:57,850 --> 00:06:03,040 trying to uncover what has taken place, where people were. But increasingly, we 59 00:06:03,040 --> 00:06:09,510 realized that we cannot really get to the places where it was most important for us 60 00:06:09,510 --> 00:06:17,491 to research. That states, when they violate human rights, or... 61 00:06:18,731 --> 00:06:23,241 T: If you want, you can... E: Okay. Thank you very much. 62 00:06:23,241 --> 00:06:26,550 Often also do two things: They close the 63 00:06:26,550 --> 00:06:35,930 area to any investigator, any human right group, or anyone that works to 64 00:06:35,930 --> 00:06:44,150 collect evidence against them. And they limit the ability of signal to come out of 65 00:06:44,150 --> 00:06:51,860 them. This is actually what happened in 2008, 2009 in Gaza. Very little signal was 66 00:06:51,860 --> 00:06:57,550 coming out of Gaza under attack by Israeli forces and this is what happened when at 67 00:06:57,550 --> 00:07:03,840 the beginning of the American drone strike campaign in Pakistan Afghanistan, they 68 00:07:03,840 --> 00:07:08,060 were effectively limiting it. So what we had to do, and this is what you're 69 00:07:08,060 --> 00:07:14,900 gonna start seeing, is to undertake archeology through looking at media and 70 00:07:14,900 --> 00:07:21,949 interpreting, really, the kind of flood of social media images that were coming our 71 00:07:21,949 --> 00:07:22,870 way. 72 00:07:40,350 --> 00:07:45,981 I'm not used to this. 73 00:07:59,460 --> 00:08:01,129 Okay. 74 00:08:21,510 --> 00:08:29,970 *applause* 75 00:08:29,970 --> 00:08:38,759 Okay, so the first example I would like to show you is work we've done with a 76 00:08:38,759 --> 00:08:45,190 Palestinian human right organization, Al Mezan, and Amnesty in Gaza in 2014. We 77 00:08:45,190 --> 00:08:50,779 could not go. We asked for permission to go to Gaza at that time. Our access was 78 00:08:50,779 --> 00:08:55,949 denied and all we had was about 70 thousand images, that we have collected 79 00:08:55,949 --> 00:09:01,039 during that day and what we had to do, really, is tell a story of one day: August 80 00:09:01,039 --> 00:09:10,170 1st 2014, morning to evening, which was the most violent, and the day where most 81 00:09:10,170 --> 00:09:15,990 Palestinian casualties were sustained. What we didn't have was the metadata on 82 00:09:15,990 --> 00:09:21,040 those clips and images, because of course they were harvested either from mainstream 83 00:09:21,040 --> 00:09:27,959 or social media and we effectively had to develop a technique to look at the bomb 84 00:09:27,959 --> 00:09:32,981 clouds themselves to look at them, and compose the architecture of the bomb cloud 85 00:09:32,981 --> 00:09:40,339 plumes in order to see and collect images that refer to the same explosion and, back 86 00:09:40,339 --> 00:09:47,029 from there, to determine the time and place of each one of those bombs. So those bomb 87 00:09:47,029 --> 00:09:51,829 clouds were for us the metadata, they were kind of like physical metadata. Here 88 00:09:51,829 --> 00:10:00,069 is how we geolocate one of the images that we found - or one of the clips that we 89 00:10:00,069 --> 00:10:09,790 found - by comparing points on a perspective of the video with what we see on a 90 00:10:09,790 --> 00:10:16,760 satellite image, located the place where the photographer was. And by cross- 91 00:10:16,760 --> 00:10:21,860 referencing three of those, of the same cloud, we managed to find the precise 92 00:10:21,860 --> 00:10:29,880 place where the bomb has finally landed. So, this is effectively a way in which you 93 00:10:29,880 --> 00:10:38,949 can reconstruct metadata, a time space location from what we call physical 94 00:10:38,949 --> 00:10:46,649 clocks, that is to say, analog things that exist within the image itself. 95 00:10:46,649 --> 00:10:54,339 However, we can see other things on this photograph, too. If we look at one of 96 00:10:54,339 --> 00:11:00,879 those images, here, we have the videographer capturing two shadow lines in 97 00:11:00,879 --> 00:11:05,470 this photograph. Just a second before closing off the camera, and what we need 98 00:11:05,470 --> 00:11:10,800 to do is to try to establish the time. If we can establish the time on that image, 99 00:11:10,800 --> 00:11:18,019 and we know that form of the cloud is that time, we can triangulate on and establish 100 00:11:18,019 --> 00:11:23,949 the times of other videos, and then move further. So, effectively by building a 3D 101 00:11:23,949 --> 00:11:29,929 model and running a Sun simulation on it, we could arrive at a very precise, 102 00:11:29,929 --> 00:11:37,170 within five minutes margin of error, time on that image. And now we know where this 103 00:11:37,170 --> 00:11:46,979 image, where this video was taken and what time it is taken in. There is, however, 104 00:11:46,979 --> 00:11:51,350 another in the satellite image that we obtained - this is a kind of a very rare 105 00:11:51,350 --> 00:11:57,630 occasion. We saw an actual bomb on the satellite image. Again, this is something... 106 00:11:57,630 --> 00:12:03,149 the satellite image has metadata. And we started looking, started hunting for that 107 00:12:03,149 --> 00:12:09,449 bomb in images from the ground. Again, if we could locate that what we see in the 108 00:12:09,449 --> 00:12:18,470 top view on from a ground view, we will be able to to start establishing times on the 109 00:12:18,470 --> 00:12:22,790 sequence. As you see, this sequence has metadata, but the metadata is wrongly set 110 00:12:22,790 --> 00:12:30,630 at around midnight. So we compose that kind of panorama of the bomb 111 00:12:30,630 --> 00:12:39,470 clouds off of the city around that time and we could identify that cloud. This is 112 00:12:39,470 --> 00:12:47,060 the cloud in side view, this is it in top view. We could find a precise location, a 113 00:12:47,060 --> 00:12:54,720 precise time, and then by confirming that is actually the same, we can move back and 114 00:12:54,720 --> 00:13:00,449 correct the digital metadata. So this is all techniques of actually establishing 115 00:13:00,449 --> 00:13:07,399 the very basic foundational stone of research time-space relations, between 116 00:13:07,399 --> 00:13:12,509 events. Here, for example, we could see in two 117 00:13:12,509 --> 00:13:17,600 different corners of the web we find those images. We can verify it's the same camera 118 00:13:17,600 --> 00:13:24,869 by seeing the same scratch on the lens. Correct the metadata, establish the time 119 00:13:24,869 --> 00:13:30,829 difference between them and now, here again, the same camera man with the same 120 00:13:30,829 --> 00:13:37,749 scratch on the lens. And now we can compose a timeline of bombs during that 121 00:13:37,749 --> 00:13:46,619 day. And after that, of course these kind of cloud atlases are a technique that was 122 00:13:46,619 --> 00:13:52,139 used by artists and by amateur meteorologists all the way from the 19th 123 00:13:52,139 --> 00:13:58,089 century on. What we did is creating that kind of archive of clouds, but here what 124 00:13:58,089 --> 00:14:04,119 you see is that we were able to convert them to information on the ground, and 125 00:14:04,119 --> 00:14:10,730 then invert the image move from cloud to city. What you see here, the model, is 126 00:14:10,730 --> 00:14:15,559 something that we call the architectural image complex. Architectural models are 127 00:14:15,559 --> 00:14:23,950 the only ways to make sense and to place those multiple images in space-time, so 128 00:14:23,950 --> 00:14:29,829 that we can navigate rather than edit them. We can navigate between one image 129 00:14:29,829 --> 00:14:34,170 and the other. What you've seen here is that on one of the images, looking so 130 00:14:34,170 --> 00:14:38,999 carefully at the bomb cloud, we start seeing two images in mid-fall. We found 131 00:14:38,999 --> 00:14:46,480 the craters, where they have landed. And, we could, for lawyers calculate the kind 132 00:14:46,480 --> 00:14:52,549 of like the destruction radius there. You would see now, again those horrific thing 133 00:14:52,549 --> 00:14:58,269 to see a bomb just split seconds before it land on the ground and would kill 16 134 00:14:58,269 --> 00:15:03,959 palestinian, an entire family. But the lawyers asked for the size of that bomb, 135 00:15:03,959 --> 00:15:11,199 in order to bring in a kind of a supply chain action on it. When we see that on 136 00:15:11,199 --> 00:15:16,369 the photo frame we can locate the photo frame within the model of the city 137 00:15:16,369 --> 00:15:22,449 and actually measure those bombs in a very precise, with a very small under ten 138 00:15:22,449 --> 00:15:27,889 percent margin of error and then go to the catalog and find exactly which bomb it was 139 00:15:27,889 --> 00:15:34,300 that landed, and that would enable activists to go after the manufacturer, 140 00:15:34,300 --> 00:15:40,410 after the policy of doing that. So again, here we are moving within the 141 00:15:40,410 --> 00:15:46,220 model, with thermodynamic specialists we look at the way the cloud is changing, in 142 00:15:46,220 --> 00:15:51,609 order to really realize we're looking at the same clouds. We're picking up now 143 00:15:51,609 --> 00:15:57,040 images and events within the city as clouds being the anchors of the 144 00:15:57,040 --> 00:16:05,519 reconstruction and that project has in fact gone later, as evidence was 145 00:16:05,519 --> 00:16:09,480 submitted to the ICC, to the International Criminal Court and was used in various 146 00:16:09,480 --> 00:16:18,829 other form of activism on the ground. And, to certain extent, might have contributed 147 00:16:18,829 --> 00:16:25,529 to a change of policy by the IDF, about the honeybell directives, that is 148 00:16:25,529 --> 00:16:31,899 something that they've enacted during that day. And the bomb cloud were also 149 00:16:31,899 --> 00:16:36,009 something that was very important, were also like memory anchors. The witnesses on the 150 00:16:36,009 --> 00:16:42,129 ground remembered and could sequence their movement according to those bomb clouds. 151 00:16:42,129 --> 00:16:47,649 So think about an element that combines and ties together material evidence, media 152 00:16:47,649 --> 00:16:57,010 evidence and memory evidence at the same time. So that's, I don't even, I'm 153 00:16:57,010 --> 00:17:03,379 basically just improvising. It's not at all the lecture I wanted to show you 154 00:17:03,379 --> 00:17:10,199 before. Here is a very recent investigation we've undertaken in 155 00:17:10,199 --> 00:17:20,240 Cameroon, and were together again with Amnesty International we were able to 156 00:17:20,240 --> 00:17:28,981 expose a secret detention center run by the Cameroonian military where 157 00:17:28,981 --> 00:17:37,960 Boko Haram prisoners or suspects of Boko Haram prisoners were actually tortured. We 158 00:17:37,960 --> 00:17:42,299 had access to people in Cameroonian prisons. It is very rare occasion, we were 159 00:17:42,299 --> 00:17:51,289 able to actually send questions back and forth and reconstruct the architecture of 160 00:17:51,289 --> 00:17:53,630 the prison and the conditions of incarceration. 161 00:17:53,630 --> 00:17:57,919 But what became very important through the questions that we continuously posed and 162 00:17:57,919 --> 00:18:06,740 continuously received from those suspects is that they've actually seen in, at the 163 00:18:06,740 --> 00:18:12,260 beginning we did not know if this was correct or not, they were obviously seeing 164 00:18:12,260 --> 00:18:18,820 people being tortured and killed outside of the detention center, but at some point 165 00:18:18,820 --> 00:18:25,850 they also confirmed seeing something different: American soldiers that were 166 00:18:25,850 --> 00:18:27,300 present on the site. 167 00:18:27,300 --> 00:18:31,591 Now, as you know the US has claimed that it has stopped rendition, and stopped 168 00:18:31,591 --> 00:18:39,110 involvement in torture, but this is something that we started very closely 169 00:18:39,110 --> 00:18:44,399 digging in to see whether we could find any traces of US soldiers and other 170 00:18:44,399 --> 00:18:53,940 European militaries involved within that, sort of, incarceration and torture of Boko 171 00:18:53,940 --> 00:19:00,940 Haram suspect. First thing that we saw, that we noticed was, and sometimes traces 172 00:19:00,940 --> 00:19:08,549 are left in the most kind of unexpected of places. A contract, an American contract 173 00:19:08,549 --> 00:19:13,140 to connect that base to the Internet. The minute that we saw that, that was put on 174 00:19:13,140 --> 00:19:19,470 the public domain we started following on Facebook and seeing some American soldiers 175 00:19:19,470 --> 00:19:25,100 actually forgot to disconnect the location tagging and you know they're kind of like 176 00:19:25,100 --> 00:19:31,640 holiday photographs. Could be very easily located onto the base. 177 00:19:31,640 --> 00:19:33,950 Again we've built a model of the base, in 178 00:19:33,950 --> 00:19:39,870 order to confirm, precisely, where each one of the photograph was taken and we can 179 00:19:39,870 --> 00:19:44,290 see that they had access to the entire base. Again, a base where people are 180 00:19:44,290 --> 00:19:55,690 executed, tortured, etc. And then tracked the unit, and as you see the site is 181 00:19:55,690 --> 00:20:01,130 actually under construction. Something that we could not believe seeing was 182 00:20:01,130 --> 00:20:07,690 American soldiers training the unit that is doing those atrocities and here in this 183 00:20:07,690 --> 00:20:14,820 almost comical film they train them in night-vision equipment by playing 184 00:20:14,820 --> 00:20:20,549 football. So they all play football in the dark with night vision and you could see 185 00:20:20,549 --> 00:20:28,480 that the involvement is very direct. So that the exposure of that base 186 00:20:28,480 --> 00:20:36,289 led to sort of a full American, at the beginning denial, always denial, then 187 00:20:36,289 --> 00:20:47,899 admission, then a full investigation by the US about these allegations. 188 00:20:47,899 --> 00:20:55,630 Another important case that we were involved with recently is the Ayotzinapa 189 00:20:55,630 --> 00:21:10,380 case. I think maybe many of you know the story of the 43 students, Mexican 190 00:21:10,380 --> 00:21:23,090 students, that were forcefully disappeared in Mexico. We were asked by the parents, 191 00:21:23,090 --> 00:21:31,210 and by other civil society group to in fact.. investigate that. 192 00:21:31,210 --> 00:21:37,389 It's one of the biggest controversies in Mexico right now, still, although it's 193 00:21:37,389 --> 00:21:45,471 three years since the disappearance. Students involved in very grassroot, left- 194 00:21:45,471 --> 00:21:53,110 wing politics enter the Intercity that was very much involved in narcos trade and 195 00:21:53,110 --> 00:21:59,539 were destroyed by the police, the military and organized crime. What we've done here 196 00:21:59,539 --> 00:22:04,779 is not really collect new evidence, but look at thousands and thousands of 197 00:22:04,779 --> 00:22:11,850 existing reports and wanting in fact to data mine them. They were, you know, 198 00:22:11,850 --> 00:22:16,210 hundreds and thousands of documents and the only way to to make sense of them was 199 00:22:16,210 --> 00:22:20,600 actually to look at relations between different events in space-time, the 200 00:22:20,600 --> 00:22:26,269 relations between all phone calls, photographs, movements of cars and gunshot 201 00:22:26,269 --> 00:22:31,759 started creating a very different picture, than the Mexican government has actually 202 00:22:31,759 --> 00:22:42,380 was willing to admit and that is that there was some local gang or local sort of 203 00:22:42,380 --> 00:22:48,380 organized crime group that was in charge of these actions. So, we created a 204 00:22:48,380 --> 00:22:56,610 platform in which we placed every named actor in space-time, a timeline in all the 205 00:22:56,610 --> 00:23:01,940 communications so that we could start seeing relation between evidence. Often 206 00:23:01,940 --> 00:23:09,491 it's not a bit of evidence in itself, it's not the casing or a gunshot that matters, 207 00:23:09,491 --> 00:23:18,399 but actually patterns, coordinations and patterns of escalations and other things 208 00:23:18,399 --> 00:23:26,200 that actually expose what was going on and we could show really a direct involvement 209 00:23:26,200 --> 00:23:35,789 between three different police forces and the military and organized crime at the 210 00:23:35,789 --> 00:23:47,100 same time, the location of all CCTV cameras that were there and removed, and somehow 211 00:23:47,100 --> 00:23:54,679 the relationship between phone calls and attacks became most clear indication of 212 00:23:54,679 --> 00:24:00,429 command and control. That these events were actually coordinated by the police. 213 00:24:00,429 --> 00:24:08,179 Here we are analyzing CCTV cameras and what they would have seen. Of course the 214 00:24:08,179 --> 00:24:14,000 state, immediately after the event erased every CCTV camera that existed, that was 215 00:24:14,000 --> 00:24:18,609 available in the city and they said or they didn't show anything. We could show 216 00:24:18,609 --> 00:24:24,784 exactly what they would have shown at that moment. 217 00:24:27,500 --> 00:24:32,220 Another element to this is, we 218 00:24:32,220 --> 00:24:38,190 go... so this is the platform, you can actually go and explore it yourself rather 219 00:24:38,190 --> 00:24:44,180 than as sort of like a work with images. As I said, this is a work with data. In 220 00:24:44,180 --> 00:24:52,480 one of the most important drawing and in fact became one of the very influential 221 00:24:52,480 --> 00:24:58,961 drawing during that, in our investigation, was a kind of a working drawing that we 222 00:24:58,961 --> 00:25:08,600 kept for ourselves, because we had to keep track of where every agent was. What was 223 00:25:08,600 --> 00:25:14,049 the relationship between them, and also the multiple narratives that were told. So 224 00:25:14,049 --> 00:25:19,610 we kind of kept a very very long drawing at the office, plotting the movement of 225 00:25:19,610 --> 00:25:25,610 different actors, until at some point we realize that what we were drawing, that 226 00:25:25,610 --> 00:25:30,540 working drawing became in fact an image of disappearance. 227 00:25:30,540 --> 00:25:34,639 Because disappearance is not about, enforced disappearance of people, is not 228 00:25:34,639 --> 00:25:41,360 only about grabbing people, killing them and hiding the bodies. Disappearance is 229 00:25:41,360 --> 00:25:47,730 also an attack on evidence. It is the continuous withdrawal and destruction of 230 00:25:47,730 --> 00:25:55,219 evidence. It is the introduction of false narratives and subterfuge. So 231 00:25:55,219 --> 00:26:03,759 disappearance is in fact a narrative form in itself and so here that drawing we 232 00:26:03,759 --> 00:26:10,220 could actually kind of like show how the state narrative, here in black, I'm not 233 00:26:10,220 --> 00:26:15,039 going to go exactly into what everything means because we lost a lot of time in 234 00:26:15,039 --> 00:26:22,529 this presentation, but these are the movement of the students according to the 235 00:26:22,529 --> 00:26:27,500 state narrative. A state narrative that is still officially holding, although it's 236 00:26:27,500 --> 00:26:35,399 being currently revised in response to many things, but including also our 237 00:26:35,399 --> 00:26:47,049 investigation. And now you would see that the victim, the survivor's narrative 238 00:26:47,049 --> 00:26:54,070 completely different. Starts... they enter the city at a completely different time. They 239 00:26:54,070 --> 00:27:06,009 move through it and the divergence between the black and the red narrative in 240 00:27:06,009 --> 00:27:12,429 fact is the space of denial and disappearance. Disappearance as an ongoing 241 00:27:12,429 --> 00:27:17,440 crime, disappearance as a crime on narrative etc. I'm gonna skip forward, 242 00:27:17,440 --> 00:27:25,919 just that you could see how the drawing is built up with another here on top the 243 00:27:25,919 --> 00:27:36,919 purple images are, the purple lines, those of the narcos and, so, each one 244 00:27:36,919 --> 00:27:45,159 tells a different story, and the multiple stories are in fact that kind of space of 245 00:27:45,159 --> 00:27:49,470 disappearance. Here you would see these are movements of police, throughout the 246 00:27:49,470 --> 00:27:56,450 city and you would see how that police force is precisely next to the students 247 00:27:56,450 --> 00:28:05,350 all throughout the attack and moving along with them. In fact, what we have finally 248 00:28:05,350 --> 00:28:16,389 done - and this is now green, is the military, etc - is built that we knew about 249 00:28:16,389 --> 00:28:31,100 the, you know, who is it ... 250 00:28:37,810 --> 00:28:40,060 So this is the complete drawing that we've 251 00:28:40,060 --> 00:28:48,259 actually printed as an enormous mural. In Mexico, murals are kind of sites of 252 00:28:48,259 --> 00:28:54,659 political pedagogy, you can think about Diego Rivera's great murals in Mexico, in 253 00:28:54,659 --> 00:28:58,999 the US, where their narratives about the history of the state and about the 254 00:28:58,999 --> 00:29:03,950 struggle of the working class. In a certain sense, we thought that this is a 255 00:29:03,950 --> 00:29:12,899 kind of a mural of the 21st century, feel like a kind of a data mural, that is 256 00:29:12,899 --> 00:29:18,370 complicated to read but its complication is in fact the image of disappearance. 257 00:29:18,370 --> 00:29:23,320 These, the entangled line and interruptions within it, is what makes 258 00:29:23,320 --> 00:29:31,584 that space.. hold on... 259 00:29:34,970 --> 00:29:40,670 exhibition... I want to show you the image of that mural in the 260 00:29:40,670 --> 00:29:49,549 space, and it has become, ever since, a kind of a site of political assembly and 261 00:29:49,549 --> 00:30:00,269 political activism, of protest, for the families and others, and it kind of shows 262 00:30:00,269 --> 00:30:08,590 for us the use of cultural and art spaces in the context of our work. Index is another 263 00:30:08,590 --> 00:30:14,039 problem in counter forensic. Very often our evidence cannot enter the very 264 00:30:14,039 --> 00:30:22,429 official spaces of state justice, they cannot. It's very rare that one can 265 00:30:22,429 --> 00:30:29,659 actually take the state, challenge the state legally in its own institutions. 266 00:30:29,659 --> 00:30:37,080 What we need to establish are alternative forums, and for us these could be public 267 00:30:37,080 --> 00:30:42,610 spaces, exhibitions, etc. I think some of you might know about the work that we've 268 00:30:42,610 --> 00:30:52,299 done on the NSU, on the Temme, or the Verfassungsschutz agent, that was suspected 269 00:30:52,299 --> 00:30:56,499 to be present in an internet cafe in Kassel during the time of a racist 270 00:30:56,499 --> 00:31:05,230 killing, and us showing that he was there. He could not have missed this event, that 271 00:31:05,230 --> 00:31:09,150 was presented in Documenta, and Documenta offered for us another very interesting 272 00:31:09,150 --> 00:31:16,809 forum. The fact it was shown there has in fact mobilized the process, 273 00:31:16,809 --> 00:31:24,710 including in a German federal investigation, and also in a Hessen 274 00:31:24,710 --> 00:31:32,090 parliamentary investigation, where different delegations from 275 00:31:32,090 --> 00:31:38,320 this parliamentary investigation came to Documenta to see it. And finally, that work 276 00:31:38,320 --> 00:31:46,470 was presented to Temme. He was forced to look at it and to comment upon it. 277 00:31:46,470 --> 00:31:52,169 I think, I should probably leave some time to question. I'm sorry about the chaotic 278 00:31:52,169 --> 00:31:58,210 presentation, but I guess this is the nature of this event, so I'm happy to have 279 00:31:58,210 --> 00:32:02,379 had at least a chance to present to you some work. Thanks for listening. 280 00:32:02,379 --> 00:32:13,909 *Applause* 281 00:32:13,909 --> 00:32:17,320 Herald: So, thank you all for a very interesting talk, despite of the 282 00:32:17,320 --> 00:32:21,629 difficulties. If you have any questions, then there are four microphones, here in 283 00:32:21,629 --> 00:32:26,980 the center aisle and two on the side. And you can line up and ask your questions, 284 00:32:26,980 --> 00:32:32,160 and first question microphone number one. Microphone 1: How much has any official 285 00:32:32,160 --> 00:32:37,500 state tried to shut down your investigations, or... 286 00:32:39,030 --> 00:32:47,919 Eyal: Well, this is... shutdown is a complicated term. First of all, 287 00:32:47,919 --> 00:32:53,159 we face interruptions initially in not being allowed access to sites. And this is 288 00:32:53,159 --> 00:32:59,809 very much the question in the West Bank and Gaza. Our investigators, when they 289 00:32:59,809 --> 00:33:05,980 land in Tel Aviv Airport, sometimes are interrogated, sometimes they're turned around. 290 00:33:05,980 --> 00:33:12,690 Nothing of that is comparable to what the Israeli state would have done if these 291 00:33:12,690 --> 00:33:19,690 were Palestinians trying to do the same thing, so to a certain extent me being an 292 00:33:19,690 --> 00:33:28,289 Israeli Jew, I'm privileged by the state. And the attempt is to use those privileges 293 00:33:28,289 --> 00:33:38,200 to undo those privileges to a certain extent. We continuously had interruptions 294 00:33:38,200 --> 00:33:44,159 from the FBI, for example, when we did the white phosphorus research, that included 295 00:33:44,159 --> 00:33:51,629 work on their attacking Fallujah, Iraq, with white phosphorus. We had some of our 296 00:33:51,629 --> 00:33:58,759 collaborators in the US's home being raided. We have, you know, 297 00:33:58,759 --> 00:34:06,679 been trolled and threatened. But, it's kind of a continuous sort of 298 00:34:06,679 --> 00:34:14,820 dance of us being to, kind of protect our staff, protect our data, and attempt to 299 00:34:14,820 --> 00:34:21,300 penetrate it. Attempts to smear us on a public domain, and feel very little 300 00:34:21,300 --> 00:34:23,730 victories sometimes. 301 00:34:26,000 --> 00:34:27,870 Herald: Microphone 4. 302 00:34:27,870 --> 00:34:32,857 4: Thank you very much for your talk. Two part question here, the first one is about 303 00:34:32,857 --> 00:34:37,126 the framework. Have you developed it special for this case, do you have it 304 00:34:37,126 --> 00:34:43,454 available, if you build new ones for each research, and the second part is: How 305 00:34:43,454 --> 00:34:49,007 do you sustain yourself financially. Eyal: Actually, it's the same question, 306 00:34:49,007 --> 00:34:53,550 because our aim is to develop new evidentiary techniques, so we kind of 307 00:34:53,550 --> 00:34:58,780 never do the same investigation twice or we never use the same methodologies twice. 308 00:34:58,780 --> 00:35:05,700 What we do, after we develop any software, is that we put it on the public domain, we 309 00:35:05,700 --> 00:35:13,120 put it as an open source code. And we, or if it is kind of techniques of more 310 00:35:13,120 --> 00:35:18,440 architectural or editing image based techniques, we have academies, we teach 311 00:35:18,440 --> 00:35:23,530 activists how to do it, so we try, whenever we work with partners on the ground, to 312 00:35:23,530 --> 00:35:30,810 leave capacity behind us. And that is also the reason... or what enables this to us is 313 00:35:30,810 --> 00:35:36,850 that we are sustained on research grants, rather than only on commissions. 314 00:35:36,850 --> 00:35:42,470 Although, you know, I mean, if a prosecutor, human right group, or any other 315 00:35:42,470 --> 00:35:47,910 civil society group would like to commission us, we would... they would pay 316 00:35:47,910 --> 00:35:52,570 for part of the investigation, but the large part of it is actually research 317 00:35:52,570 --> 00:36:00,930 grants that translated into open source stuff, and the investigations are being 318 00:36:00,930 --> 00:36:04,390 put in the public domain. It's kind of, when you look at our videos, they're a little 319 00:36:04,390 --> 00:36:08,441 bit like cooking programs, because they both tell you what we find, and they tell 320 00:36:08,441 --> 00:36:13,190 you exactly how to do it. It's kind of, take you step-by-step, this is what you do 321 00:36:13,190 --> 00:36:18,290 here, then that, then this, etc. 322 00:36:18,290 --> 00:36:19,920 Herald: Microphone 4 323 00:36:19,920 --> 00:36:23,790 4: Thank you so much for your work on this. My question is like, how would any 324 00:36:23,790 --> 00:36:29,760 of us be able to get involved in this, support you in one way or another. 325 00:36:29,760 --> 00:36:38,590 Eyal: We, in fact, we are now about 15 architects, coders, and filmmakers and we 326 00:36:38,590 --> 00:36:45,260 are recruiting because we're growing. I will stay here for the day, so anyone that 327 00:36:45,260 --> 00:36:50,530 wants to come and work with us in London or remotely, I'll be delighted to speak to 328 00:36:50,530 --> 00:36:56,210 you. Herald: Microphone 4 329 00:36:58,960 --> 00:37:05,070 4: I have a question about which tools, which techniques and tools do you use to 330 00:37:05,070 --> 00:37:11,600 perform the 3D reconstruction, and if you have partnered with any kind of company 331 00:37:11,600 --> 00:37:17,820 that already has *inaudible* and already does *inaudible* 3D *inaudible* as a 332 00:37:17,820 --> 00:37:21,960 baseline. Eyal: We started doing now 333 00:37:21,960 --> 00:37:29,580 photogrammetry, as 3D reconstruction from existing open source images. So 334 00:37:29,580 --> 00:37:35,580 imagine, you know, a place in Syria, let's say, that has been photographed or video-ed 335 00:37:35,580 --> 00:37:41,230 by many users. We are able to reconstruct it. In fact this is one of 336 00:37:41,230 --> 00:37:45,961 the techniques we use in order to identify the gas attack on Khan Sheikhun in 337 00:37:45,961 --> 00:37:52,120 Syria by the regime forces. Reconstructing precisely, to the millimeter, the shape of 338 00:37:52,120 --> 00:37:59,480 the crater. And we were able to reconstruct from it the level of explosives etc, and 339 00:37:59,480 --> 00:38:06,230 that they were fitting only that particular rocket. We don't really work 340 00:38:06,230 --> 00:38:12,810 together with companies, we try to take existing softwares and kind of 341 00:38:12,810 --> 00:38:20,250 adjust them to our aim. But initially, what I want to leave you with, is 342 00:38:20,250 --> 00:38:23,400 the question of why architecture is really 343 00:38:23,400 --> 00:38:30,080 important here. In a situation when you don't have only like two images of the 344 00:38:30,080 --> 00:38:36,900 scene, let's say police brutality or an attack on a city etc, but you have 70,000 345 00:38:36,900 --> 00:38:42,500 and you need to cross-reference them and you need to place them within a space, the 346 00:38:42,500 --> 00:38:49,000 only way to do it is in architectural models. Architecture is like the optical 347 00:38:49,000 --> 00:38:54,760 device that allows us to sync up and locate, you know, those cameras that are 348 00:38:54,760 --> 00:39:02,400 in space and moving in space. So, it is really the necessity of work 349 00:39:02,400 --> 00:39:11,370 of architects, filmmakers, and coders is fundamental, because space replaces the 350 00:39:11,370 --> 00:39:16,980 kind of modernist montage as a relation to images. Montage is the edits in film, that 351 00:39:16,980 --> 00:39:22,230 is kind of, you know, the basic of cinema, of political cinema, the dialectic montage, 352 00:39:22,230 --> 00:39:25,100 if you like. You splice film and put it together. 353 00:39:25,100 --> 00:39:31,030 That makes no sense for us, because we need to move within space, pick up one 354 00:39:31,030 --> 00:39:35,530 film, not to cut it, we never cut the films that we have, we just leave them 355 00:39:35,530 --> 00:39:41,380 within the model in the full duration, but the investigator can move and navigate in 356 00:39:41,380 --> 00:39:46,900 space and time between them, and as I showed you in a Mexico case, you know, 357 00:39:46,900 --> 00:39:53,910 these are like tens and tens of thousands of data points that create kind of 358 00:39:53,910 --> 00:39:59,860 intersections, between data image and architecture, where the story starts 359 00:39:59,860 --> 00:40:11,850 to unfold at all. So yeah. Herald: More questions. Microphone 3. 360 00:40:11,850 --> 00:40:17,580 3: When you publish the videos and the time and place from which they were taken, 361 00:40:17,580 --> 00:40:22,690 how do you ensure that you're not putting in danger the people who took the video. 362 00:40:22,690 --> 00:40:29,310 Eyal: Yeah, this is a really good question. We..., the work to sync up those 363 00:40:29,310 --> 00:40:36,410 70,000 images from Gaza, where... took us a year. Think about it, a year... we're 364 00:40:36,410 --> 00:40:42,220 working a year on one day. That's about the right kind of ratio in forensic time that 365 00:40:42,220 --> 00:40:47,830 we are operating within it. What protects people during war, during when 366 00:40:47,830 --> 00:40:55,810 they will do it, we'll never place their location, but months after the conflict it 367 00:40:55,810 --> 00:41:04,190 was deemed by our partners in Palestine and by our partners in Amnesty, that this 368 00:41:04,190 --> 00:41:11,760 is safe to do without going back to each source and in fact asking them. We would 369 00:41:11,760 --> 00:41:16,740 never do it in real time, though. Herald: Microphone 4 370 00:41:16,740 --> 00:41:27,400 4: Yes, the work you do strikes me as very similar to what bellingcat do, so can you 371 00:41:27,400 --> 00:41:29,520 comment on how forensic architecture compares to bellingcat? 372 00:41:29,520 --> 00:41:33,240 Eyal: No, we work a lot with bellingcat and Elliot, I mean some some of our 373 00:41:33,240 --> 00:41:42,960 projects are together, I guess that our... the difference is not, we engage 374 00:41:42,960 --> 00:41:52,550 more in sort of big environment and kind of like data analysis, from many sort of 375 00:41:52,550 --> 00:41:58,560 data points, where architecture, or architectural models, are kind of the arena 376 00:41:58,560 --> 00:42:04,590 that holds and cross-reference all those images together. The overlap in our work, 377 00:42:04,590 --> 00:42:10,140 really, is a kind of image identification: What do we see, where the image is located 378 00:42:10,140 --> 00:42:22,760 etc, and on these issues we work with them together. We tend to work more against 379 00:42:22,760 --> 00:42:32,740 states', western states', militaries, holding them to account, we feel, is that these 380 00:42:32,740 --> 00:42:39,390 techniques are actually much more useful directed at the British, American, Israeli 381 00:42:39,390 --> 00:42:48,030 militaries and that we are able also to draw responses that are effective in these 382 00:42:48,030 --> 00:42:55,230 fields and I guess bellingcat has slightly different sort of field in which 383 00:42:55,230 --> 00:42:58,695 they work. Herald: Okay, I think with that, that 384 00:42:58,695 --> 00:43:01,825 would be our last question. So again, a big round of applause to Eyal for a great 385 00:43:01,825 --> 00:43:03,203 talk. Eyal: Thank you. 386 00:43:03,203 --> 00:43:05,714 Herald: Thank you very much. 387 00:43:05,714 --> 00:43:13,330 *Applause* 388 00:43:13,330 --> 00:43:34,421 *34c3 outro*