#17 - Terror Element: The Fragility of Truth in Forensic Science

Welcome to the FramerFrame podcast in which we take you behind the scenes of the art platform FramerFrame in Amsterdam. In today's episode, a recorded lecture during the symposium's synthetic vision images of power. The lecture by Anna Engelhardt and Marc Sinkiewicz is about their work in the exhibition really currently on show at FramerFrame. The film, terror element, is a hybrid documentary exploring the investigative method and the fallibility of truth. Anna Engelhardt is a video artist and writer, Marc Sinkiewicz is an interdisciplinary researcher and artist born in Belarus. The exhibition is open for visitors until September 29, 2024. So hi everyone, I'm Marc Sinkiewicz, this is Anna Engelhardt, so I had to wear a mask because I need to hide my face, so I don't appear in the pictures, but since we don't have them taken, I can be resorted and can breathe, which is amazing. So, I'm a research-based artist, I focus on taking emerging trends in post-soviet infrastructural social landscape, but more generally I work with the image production and infrastructural colonialism and monstrosity, and I do PhD in anthropology at the New York Warsaw. Anna Engelhardt is a research-based media artist and writer, and her practice she examines war as a technology, looking at the hardware and software behind Russian invasion. She is interested in topics from military cybernetics to cyber warfare and conducting investigations that take multiple forms, multiple media, drawing from her education in forensic architecture, from Goldsmiths. So in our work, we basically let me reflect the core methodology in our practice, it combines first of all open-source investigations in peripheral narratives, and we work with Russian colonial violence at all, we've been working with it for quite a while, long before the escalation of Russian invasion to Ukraine, and we have been critically resinking our practice to address the change in context since the full-scale war. And before 2022, Russian expansion was undertaken a rather like covert operation, where every aspect of it was really denied, so we had to work a lot against masking, against hiding, and the task was to prove that Russian soldiers are actually there, that they invaded the country, it was the task in itself, but since the full-scale invasion, the military strategy has changed and now Russia publicly takes pride in what is going on similar to Israel at the moment. And what is the new role of the investigatory practice in such an environment, what does it mean to reconstruct a crime, to address a crime, where perpetrator is just proud of it. And so for this exhibition framework, we made a new work, which is called Terror Element, and it's a hybrid documentary about the investigative method and the fallibility of truth. It's set in the present day, it follows Nina, a reclusive forensic expert who spends all of her time in crime laboratory, like many of us, and not in laboratory, but just at work. When she rediscoveres a videotape by her mother, made about the notorious series of explosions in Russia that took place in 1999, she also has to confront the uncomfortable truths behind her face in forensic science, as a forensic scientist. So the work is comprised of CGI and archival footage, and it pieces together those contradictory statements that were presented by Russian authorities. And it focuses specifically on an unknown sentence that was found at the pompsides. In this attack was widely televised, and the investigation was also widely televised. And forensic laboratories became a stage for the sort of real-life crime fiction that was unfolded in front of everyone. And moreover, those explosions became a pretext for the second invasion of Chechnya and start with the second Chechnya war. As Nina's investigation spirals further into uncertainty, her own methods become a part of why they're conspiracy. So in the work we're specifically focused on the chemical analysis, but I guess we can say that the topic of the work in general is the fallibility of forensic evidence. And I want to start quite far in that with the term that all of us have heard of, which is deduction. And I don't know how many of you have had classical logic in their ventulars, maybe your musters. But if you had classical logic, you would know the deduction, deduction is something that is a statement that is made from two general factual, infallible premises. And from those two premises, you conclude something particular. So let's say premise one, all men are mortal, premise two, cicadas is a man, conclusion, cicadas is mortal, investigators on the other hand, who claim to deal with a deductive method, deal with, and this is very important, probable interferences based on observations. There is always a possibility of something being true, but we do not know it for a fact. Let's imagine there is a room and there is a window broken and there is a shards of glass inside the room. A Sherlock Holmes would come and say, well, it is obvious that the break into the room happened from the outside because the shards of glass are inside the room and not outside the room. But this is not deduction. To say something is a deduction is to be 100% certain, to have it as probable truth. Whereas in this case, there is still a slight chance, a small possibility that someone did just place the shards of glass inside the room. And there are also unknown possibilities like this. So originally when the forensic investigation, forensic science was born, the deductive method was claimed to be used purely to make detectives appear more trustworthy than they actually were. So starting with this, what is the possibility, what is the quantified chance of something being true when we approach it from a perspective of forensic science? And when we talk about fallibility in forensic science, discovering all the very interesting things, when forensic methods are being tested today in scientific journals, in the examination system, there is not just yes or no system. I do not know, dated by the expert, is marked as a correct answer. I want you to do a bit of a mental experiment and imagine your passing state exams in school. And let's say you're passing math and you have an option of stating in your state exam that you do not know all the answers, and you would get a 100%. Why? Yes, I don't have to explain why in a state exam, this is not a good idea. This is not a good idea in the forensic science either, because what forensic experts do is they state inconclusive results as responses, but when those inconclusive responses are marked as incorrect, instead of correct. The overall error rate that is usually marked around 1% skyrockets to 52%. 52% of inconclusive responses at the time where forensic technologies are being tested are never transferred into the court case scenarios. In the court, the inconclusive response is extremely unlikely. This is due, and why do forensic experts not state that they might not know something? This has to do with what is called CSI effect. I don't know if you know CSI as a series, but there is a cultural phenomena where CSI effect is literally referring to the effect where the capabilities of forensic science are overstated. I guess we all know the example of the super-resolution or hyper-resolution, where we can endlessly zoom into the same picture to uncover new and new layers of information. This is true for the forensic science as a total. People who are brought up in the CSI effect culture are then made to believe that they have God-like capabilities of discovering information. This allows investigators to make assumptions based on miniscule in sufficient bits of material evidence and information taken out of context. And as one of the examples, I want to show a specific discipline that was born like this that is called blood splatter analysis. The person who made blood splatter analysis had nothing to do with physics. He had no physics education, no mathematics background. He loved cutting pigs, and he loved covering his female interns in blood of those pigs. He called himself Sherlock Holmes, co-authored a fiction about himself, where he appears with a pipe on the cover of the book. It all started, the blood splatter analysis became coded as a scientific discipline after a case where he testified in court where in a white code and delivering an entire performance where he was dropping a liquid stating that he can predict where the liquid is going to drop. And the problem with that is if you are familiar with the fluid dynamics, it's an extremely hard discipline in terms of the prediction of where something is going to land. When blood exits the body of a human, it becomes mixed with air, so making assumptions about where exactly which exact pattern is going to result in is almost impossible. What blood splatter experts do is they project their assumptions on what is almost functioning as a Rorschach test and then testifying court stating that it's unfalable truth that they can send behind. So this is one of the examples of CSI, in fact, having an impact in the forensic science. This has been heavily criticized, as you may imagine. And specifically when there was a report published in the US regarding a statement a phrase that is called reasonable degree of scientific accuracy. The interferences, a lot of things depend precisely on the method that is used and one of the things is that methods have a right degree of being based on science. One of the things is that some of the forensic disciplines like firearm analysis, for example, you might have heard of it, so where they could detect the claim, they could detect if the bullet was shot from the same gun or not based on the scratches that are left on the bullet. So this particular bench was using a lot of the formulation of reasonable degree of scientific accuracy. This was presented in court and this was recognized to be so misleading that this formulation was officially banned in 2016, but only in 2016. And instead, this was substituted by some more complex formulation that essentially expressed pretty much the sensing. And it points out in the direction of a lot of aspects that are not done in line with what we consider to be a proper scientific process. So one of the things is that many of the testing of the forensic methods are done not by scientists and not in the scientific laboratory, not according to scientific methodologies, but they are done by the producers of, either those testers, if you're talking about chemical analysis, even prisonisers. So you can imagine that people who work on the product, they have sort of a bias. They inclined to give the information that would allow the product to sell better. So instead of scientists, the forensic people should be rather called experts or practitioners instead. And it also points out to a whole range of problems with accessing the information about how the analysis is done. So one of the things is access to the evidence. For example, only the accusation site, only the police has the priority of accessing the evidence. It's only them, the defense site, either never has the access to it, or it has an access while being supervised by the police. One of the very crucial things is that in scientific journals that deal with the forensic, there is impossibility to compare data that is presented there because one of the police policies is the destruction of evidence. And in many cases, it's either not stored and destroyed directly after the case has been closed or in certain times, like in two years, because eventually they run out of storage. They just close the case and destroy the evidence, which doesn't allow to retest and doesn't allow to compare the results from different laboratories, which lives in sort of like a fork. And of course, the laboratories, they are a part of the police department always, which makes the forensic and practitioners see themselves as a part of the police department. They just consider themselves a part of the team. And so this information, while the evidence is usually collected by the police, who are not themselves scientists, this evidence is collected in various ways with many violations. Then it gets transferred to the laboratory. In the laboratory, the scientists have access to it. But then again, it's where they're all ends. And when the information travels to the court, it's being presented again by not scientists, but by curious or just like police representatives, and they could roleplay as scientists by varying like white suits, for example, and the mediation, the technological mediation here plays a key role, because they outsource their responsibility from themselves often to the testers or to the process of chemical analysis, which types into the other branches of science. So which allows them very often to present absolutely contradictory information, as was in the case with the explosions, where the same sample can be claimed to be either explosive or not explosive, completely based on the scientific process of chemical analysis. So yeah, and instead of being the process of actual scientific exploration, it becomes, in the tools, becomes sort of aid for the police in the process of law enforcement. And here we come to the strand of forensics that we think is directly connected to our today's symposium, which is image forensics. Image forensics operate on a very interesting assumption. Here, I'm going to read a Q&A from a forensic tool Cognitaq that is used extensively. So let's see how they describe their software. And here I want you to point your attention towards the idea of image restoration. There's an inherent belief that there is an hidden ideal image within each image. And one can use certain tools to access this ideal image, which is always preserved. It is not that a new image is created in the process. It's the ideal image is reconstructed, brought to the surface, and its imperfection, let's say, blower is removed. Images that require the information quality improvement have been afflicted by some physical degradations, otherwise, they would be perfect, e.g. too little or too much image formation energy, and/or optical blower, et cetera. Just as it is in medicine, the success of treating the afflicted patient, the most important part of successful treatment, is to diagnose the disease. In a practical environment, one can think of triage process when perhaps several degradation sources are present. Once the triage is properly established, the practitioner would look for the appropriate software tools in his or her arsenal, the mastery of these tools then becomes equivalent to formulating the optimal treatment plan. Anyway, they are implying you should also buy from their website a course to learn how to use their software. Another question that is also fantastic, can context software enhance any license plate? This is, for instance, what they mean with license plate enhancement. If anyone like us have any experience with Photoshop, or on that matter, a generative AI, I don't have to explain what is the issue of using this as an evidence in corn after it has been treated. And this is not the problem, let's say, only with license plates, and this is one of the examples once again of the CSI effect, right? There is a series in the 90s that everyone believes that we can enhance image and there is some ideal image, etc., etc. And then there is a generation of people who genuinely believe those ideas and who make tools match the expectations that are put out there. Forensic methods also go in and out of fashion. For instance, in the first half of the 27th century in the 1950s, one of the most popular forensic methods was bone analysis to determine the race of the person who's been searched. Today, obviously, the fashion is AI. The image analysis tools that we're showing are now being augmented by AI to go into the trials for better image analysis. But image analysis does not just function as an object, let's say, to be treated or triaged. But it also functions as an appeal of a scientific tool. There is a certain promise, as we all know here, in the forensics, it is quite explicitly realized I'm going to read you out loud the introduction to the forensic investigation as it was. On the basis of Article I of the Criminal Procedure Court of Blah, blah, blah. The participants of the examination were notified about the use of a Sony video camera during the examination. Criminal persecutors Volkov UB, Adaptive AG, Minolta 700 camera with Minolta RF Zoom 2485 lens and Super 36 by 400 product, 12 by 400 for the graphic film, Aga Karakosov in place of the state investigation department of RF GPU video camera Canon VCX45 camera, Nikon F50, 24, 1, 2, 0 camera Fuji Super 100 laser range funder, Laker, etc, etc, Lisco goes on. Once again, as we love doing the mental exercises, I would suggest that if we, as experimental documentary filmmakers, make a disclaimer in the beginning of our film, listing all the hardware that we're using in the attempt to claim that our film is truthful. And this is not something that is purely coincidence as we, who fully successfully demonstrated, tools in forensic science are an important part of the performance of science. Tools are the science in the forensic science. But these tools do not just legitimize the industry itself. These tools legitimize frequently violent processes that they are engaged in. And for that, one of the best examples would be the light detectors. Light detectors have officially been proven to be fake science. They are, as you might imagine, are still used today. For instance, in UK, they're extremely common. You can be released as a terrorist earlier than your sentence if you don't pass the light detector test, even though light detector tests are proven to be unscientific. And they're actually manufactured to provide, because they measure how nervous you are. I mean, if you put us in front of light detectors right now, we'd definitely be lying. So it's a tool that is hardwired to produce false positives, to convict as many people as possible. And this is also why they love so many pixelated images, because the more degraded the quality of the image is, the easier it is to interpret the way you want. The thing with the polygraph, and this is why I wanted to use it as an conclusion, is polygraphs were extensively used by US military in the tortures in Guantanamo prison. And light detectors are still used today in the same fashion. They are meant to bring validity to a process of torture. They're meant to color torture and inhumane treatment in a way that appears to be a scientific process, process of gathering information rather than human rights violation. So unlike science forensic often works against the bad guy, they have, they are goal-oriented pretty much. And those of you who teach, you might know how students sometimes do like backwards engineering, they have a statement that they want to prove, and then they arrange the arguments. So this is pretty much how it happens and the extent to which people without scientific training tend to believe forensic science evidence, it's expected to bring this sort of neutrality, fairness, accuracy, but precisely this mediation through tools and methods allows to do forensic precisely the opposite, to add false legitimacy to illegitimate convictions. And many forensic methods have no use besides law enforcement. forensic methods do not only enable carceral harm, they also launder and legitimize it by clogging it again under this allure of science, under the neutrality, under the scientific image. And as you know, in the recent years we saw the explosion of prison population and this is connected not with the fact that people in fact started committing more crimes or became worse, but connected to the increased efficiency of forensic tools and forensic methods in securing convictions. And the ability to secure convictions and to prove the linkage between a suspect and the crime is in itself taking place of the validity of the matter. And yeah, we'll finish here. Thank you for the attention. [BLANK_AUDIO]