Narcissurveilance

Build - Crop

    It was walking the streets of Melbourne’s CBD that the idea came to me. I had spent most of my adult life interacting with machines; first as a programmer/ operator, then as a designer, builder and commissioning engineer. Walking down La Trobe St. Melbourne, I discovered for the first time, a new generation of parking meters; ‘smart’ parking meters (2017), offering a multitude of ways to pay,   making everyone’s life so much better. To my amusement, the meter had an illuminated scrolling banner sign above it, declaring it was a parking meter and informing the public to park in ‘bays’, reminding me of some sort of fun park ride. I had found my physical structure, I had found my machine. I had already made a machine that looked similar to this, but I never gave it any consideration. This previous machine was actually four identical machines, standing side by side, in a remote rail yard in the Pilbara, W.A. Their function - to test brake cylinders on iron ore wagons for leaks and other defects before being sent back out the tracks to the mine. These machines were pure function, designed to withstand their harsh environment. Yet on reflection, there was some beauty in their polished stainless-steel aesthetic and repetition that I had not acknowledged at the time. This new machine would take its cue from the smart parking meters; yet will not accept any forms of payment. This machine will not allow you to ‘interface’ with it via a pushbutton or touch screen; it will transfer the control away from the public and place it in the hands of the machine. This machine would act as a type of autonomous identification checking station. 

IMG_7024.jpg
Melbourne's new 'Smart' parking meter, 2018
DSC_3184.jpg
Brake Cylinder Leak Test Machine x4, Pilbara, 2014

    The Oxford dictionary defines a ‘node’ as ‘a point in a network or diagram at which lines or pathways intersect or branch’; in physics,‘a point at which the amplitude of vibration in a standing wave system is zero’; In both instances, a node is a stationary point surrounded by movement, but it is also part of a larger collection of points. American theorist Benjamin Bratton argues that, in reference to sensors, ‘instead of seeing the various species of contemporary computational technologies as so many different genres of machine, spinning out on their own, we should instead see them as forming the body of an accidental megastructure’ (Bratton, 2014). The megastructure Bratton refers to is the topology of all things smart, that sit in that the space between user and function. This machine I was proposing was ‘smart’ and I wanted to suggest it was part of a much larger collection of devices, part of Bratton’s megastructure. Yet, the data this machine was going to collect is of no value without a greater context, a time and location. Each subject’s collected face would be time-stamped and recorded against the sensors' geographical coordinate. Titled The Geo Facial Node, this sensor will capture and collect facial images of the public and interpret them as donors to the fictional Smart Urban Infrastructure. These donors will be informed of the process but not be able to control it. 

    This sensor is both an idea from the future and something from the current day. In its physical form, as a street located device, it stands as a very totalitarian symbol, one that would not be accepted in our modern culture. Yet in its philosophical form, it already exists. It exists in the fact we accept CCTV technologies on our streets, yet we really do not have a full appreciation of how that collected information is being used. The Guidelines to surveillance and privacy in the Victorian public sector 2017state that ‘The rights of individuals to access their personal information should be respected’(Government, 2017), yet there is no means to actually quantify who might have that information. It can be assumed, in certain locations, who might own the CCTV infrastructure (such as those found on RMIT’s or any other University campus) but how can the public hope to find the owner of similar equipment in the public realm? What is the process for accessing the information? The public might have civil rights to protect them but in the digital age knowing your rights have been violated will prove to be the real challenge. 

RMIT_CCTV_1.jpg
CCTV's Located on RMIT's City Campus, 2018

    Playing up to this idea of surveillance and acting as what I hoped would be the drawcard for the work, was the public's narcissism. Art theorist Professor Rosalind Krauss wrote, in reference to video art, that ‘video is capable of recording and transmitting at the same time – producing instant feedback. The body is therefore as it were centered between two machines that are the opening and closing of a parenthesis. The first if these is the camera; the second the monitor, which re-projects the performer’s image with the immediacy of a mirror’(Krauss, 1976).This instant feedback was what the audience enjoyed so much in the earlier work The Zero Monument, yet the technology differences between this work and what Krauss was referring to are vast. It would be interesting to see if this feedback excitement could be replicated in a slightly delayed and not such a predictable situation. 

The GeoFacial Node would be a three-part installation, the Node itself, a larger screen that would playback the captured faces – publicly naming shaming them, and a third element, a sound box playing a text to voice version of issues related to facial recognition software.

It was during the disassembly of the work The Zero Monument early 2017, that I started to really think about the aluminium extrusions, that made up the internal frame of the work, as a material that deserved to be exposed, not hidden, in a future artwork. Olafur Eliasson, an Icelandic-Danish artist best known for producing some very refined artworks out of steel and mirrored surfaces, produced a work that I find the rawness of quite rewarding, that is Waterfall (Eliasson, 2004). Eliasson had made several waterfalls over his career but it is this one from 2004 that pays homage to the structural materials, the scaffolding that makes up the supporting framework. In particular, I note that the scaffolding he used is not new, it is soiled with stickers and paint from previous use. This gives the work, which is abstracted idea of natural beauty, an opposing, gritty, man-made element. The aluminium extrusions from The Zero Monument were salvaged from a factory bin and showed the signs of age, of a life and purpose bound by practicalities and function. It seemed fitting that they should be reworked into the frame structure for The GeoFacial Node and the sound box. There was also something quite pleasurable in working with these extrusions; once they were cut to size they could be fastened together with some speed that gave a real sense of progress (as opposed to learning to computer program).

The Zero Monument, 2016, in a state of dis-assembly

    Cladding the externals of the GeoFacial Node and the sound box was done with the acrylic panels from the Unnamed Laneway. These panels could be cut to size and inserted into the aluminum extrusions; not only did they go part way to replicating the polished stainless-steel aesthetic, but it also provided a duality to the camera/ screen machine mirror by encapsulating it in mirrored surface. I had intended to hide the screen behind a one-way-mirror, but the practicalities of screen illumination and cost resulted in a cutout allowing access to the screen. The mirrored panels were now two years old and had been used for numerous tests by other artist at RMIT, the remnants of these tests and the general collection of dirt and grime had taken their toll on the reflective surface. This surface history on the mirrors and the aluminum extrusions enhanced the utilitarian aesthetic I was attempting to achieve. 

GeoFacial Node, early concept model

    Michael Stevenson, the New Zealand born artist, produced a machine of his own in 2006. His machine titled The Fountain of Prosperity (Answers to Some Questions About Bananas)(Stevenson, 2006), was a functional replica of the world’s first economic computer, a hydro-mechanical marvel using fluids in various chambers to represent variations to the fiscal balance of an economy. His main focus was on the use of the machine in the Central Bank of Guatemala and the failed western takeover of the tropics. What grabs me about this work, beyond the fact that here is an artist presenting a machine as art, is the manner in which he chose to let the condition of the machine deteriorate ‘reducing it to a decrepit, ruinous, economic state’ (Stevenson, 2006), using the material state as a metaphor for the economic condition of Guatemala. What I was attempting to achieve in my choice of materials was to start at some point in the past and build in a history to the machine through the material aged appearance, suggesting that this technology and equipment had been present, in the public space, for some time already and was only now being noticed.

    This work was originally conceived to be situated in the area of the footpath generally reserved for urban infrastructure, human- system interfaces, street signage, the electrical grid and the occasional seat. However, the practicalities of power and security as well as a certain unease, on my behalf, as to how the public would respond, forced to me look elsewhere in the public space for a possible location. The location I choose in the end was Melbourne City’s Testing Grounds, a public space for artists to ‘test’ their ideas. This wasn’t the most adventurous of locations, but it certainly addressed issues of services and security, and perhaps the patrons of Testing Grounds would be a little more accepting of having their image captured. And this was very much a work that needed testing, both from a conceptual perspective but also from a technical one. I was working beyond my capabilities and with technology I have never encountered before so there was no guarantee it was actually going to work.   

A site visit to Testing Grounds showing the super structure, the information booth and the main walkway.

    The location within Testing Grounds needed to have a reasonable level of daily traffic to make this worthwhile. Thankfully, there is a coffee cart that operates every weekday morning, outside of Testing Groundsnormal hours, giving opportunity to expose the work to the general public as well as art patrons. These coffee drinkers funneled through the space on a well-trodden path past the information booth, so this was my focus. One of my concerns with my early concept was the physically stability of the unit; without fastening it to the ground, could it be stable enough not to fall over? Even with a heavy ballast, it could still be pushed over. Standing at the information booth the solution became obvious – the whole area is covered with what is referred to as the superstructure, a galvanised steel frame made from 90mm square section material. The only issue was that at the base of each vertical section was a large concrete foundation, hampering the idea that The GeoFacial Node’s frame could be floor standing and be fastened back to the superstructure for support. To make use of this frame and without modifying what I had already made, I had to abstract the idea of the parking meter. Rather than standing like a tall tombstone, The GeoFacial Node’s frame could be attached to the frame horizontally and be partially cantilevered out to allow the participating public to approach the sensors screen. 

    Maintaining a focus on the area directly in front of the information booth, the three works were arranged in the corners of this area. The GeoFacial Node was mounted, as previously described, onto the superstructure; diagonally opposite was the monitor – mounted up above the information counter, onto the booth itself. . Directly in front of the monitor was the sound box, closing out the area to the rest to the grounds. I was attempting to produce a small area where the public would be enclosed by the work, on three sides at least, and generate a function driven space not unlike a bus or train station – ubiquitous in nature and blurring the line between artwork and utilitarian objects.

    The American artist Tom Sachs is someone who produces works that blurs that line. Often his work has an element of parody about it, such as Chanel Guillotine (Breakfast Nook)’(Sachs, 1998), where the title alone gives it away. But it is his short film titled Incinolet (Sachs, 2008-2011)that blurs that line so well; the film is set on the bridge of a U.S. aircraft carrier and features a young woman making use of a new type of toilet; a toilet that incinerates the waste rather than flushing it. It is only through the presentation of the work that it becomes clear that it is a fictional object – an artwork, and not a utilitarian object. Much like Sachs, I was striving for the same end effect, using my background in machine design to make The GeoFacial Node as utilitarian as possible, yet, counter it with something absurd. 

The GeoFacial Node, 2018, Initial Build

    The work was installed for a 2-week period and collected over 2000 faces. Playing up to the idea of the new ‘smart’ parking meters, a scrolling banner sign was installed above The GeoFacial Node that encouraged the public to ‘Scan Your Biometric Here’. As the public approached the screen of the node, a sensor would detect their presence and turn on a LED strip light, illuminating their face. A hidden webcam would capture their face automatically and display it back to them on the screen along with some information pertaining to their image, date, filename and geolocation. The public had no control over when or what image was captured, this was under the control of the software. Once a face had been captured and stored in the local hard drive, the containing folder was synchronised with another computer operating the monitor attached to the information booth. The second computer was running a software package called VVVV; here, the images were overlaid with text and sent to the monitor, displaying two scrolling sets of images that would intermittently pause on an individual face and accuse them of a crime randomly taken from the Victorian Police Offences Recorded by Offence Codelist (Police, 2004-2015). Trevor Paglen is an American Artist whose work often questions ideas of state surveillance and monitoring in the digital age; Code Names of the Surveillance State (Paglen, 2010) is text work, projecting the names of U.S. Government surveillance programs onto the outside of public buildings. Paglen was attempting bring awareness of the extent of modern day surveillance, the text I was overlaying onto peoples faces, as an accusation, was to personalise the work in a confronting manner but also to hint at the impersonal manner in which a computer and artificial intelligence can accuse you of a crime.

    The third element of this installation was a sound box playing an audio work comprising a text-to-speech extract from the Electronic Frontier Foundation’s (EFF) white paper Face Off (Lynch, 2018), mixed in with sound bites of the torture scenes taken from the film 1984 (Anderson, 1956). This audio track was highlighting the dehumanisation of facial recognition software and its applications through the use of text to speech; complimenting the content, a warning of current and predicted failings of the technology. The scream sound bites warn of a future where the technology is not longer working for us. 

The GeoFacial Node, 2018, Testing Grounds, Melbourne

    As was expected, the work was full of technical issues when it was first installed, programs would not run, computers would not communicate with each other and covers needed to be made to protect the work from the rain. I had each element functional as individual works but bringing them together is always problematic. In an effort to extend my Linux knowledge, I was determined to use a Linux based Raspberry PI computer to control the facial detection and collection software. If this worked, it opened up other possibilities in the future for a battery power installation somewhere. However, in the rush to get it working during the installation it was decided to replace the Raspberry PI computer with an old Windows computer. Not ideal but at least the operating system would match the computer behind the naming/ shaming monitor and this is where I felt most comfortable. Once the complete installation was up and running, another glitch became apparent; the facial detection software was insistent there was a ‘face’ in the pile of bricks stacked to the side of the Information building. This meant that pretty quickly a collection of the same image (the brick pile) was growing by an image a second. A little tuning of the what the Haar-cascade was actually identifying as a face made the results a lot better; although not perfect, perhaps due to shifting light, there are big blocks of random images (brick piles, a step, etc.) throughout the collection. I think this glitch is representative of issues that other applications of this type might encounter, issues that might affect the decisions artificial intelligence makes. Jennifer Lynch, from the EFF, highlights the accuracy issues with facial recognition software by arguing that significant errors result from ‘photographs that are compared to one another containing different lighting, shadows, backgrounds, poses, or expressions’(Lynch, 2018). Not to suggest that you might be mistaken for a pile of bricks but even in this situation, where the camera is fixed, the lighting is somewhat artificial and therefore controlled, the results were very mixed.

    The arrangement of the three pieces did bookend the space directly in front of the information booth. The GeoFacial Node capture device played well against the monitor mounted in the wall of the booth, but it was the soundbox that blended into the surrounds. Identical to The GeoFacial Node aesthetically, aluminium extrusion and mirror panels, but with no exposed elements, it just sat mounted to the super-structure, diffusing the audio track, that was long enough to trigger more interest with each passing..

The GeoFacial Node, 2018, Screen Capture

    From the faces that were collected it is obvious that most people approached the work with some curiosity, then as they start to see themselves appear on the screen, they play up to the situation, smiling, pulling faces and generally being silly. Each of the captured faces was timestamped to the millisecond and was recorded using the timestamp as file name. As intended, there was no way of looking at both the GeoFacial Node’s screen and the other monitor simultaneously. This separation really highlighted the audience’s frustrations at not being in total control, ‘I am here on one screen but not the other’. The public name and shame monitor would cycle through the images in order, so even if your face has just been captured, it could take quite some time for it to appear on the monitor, even longer for any accusation to be made.

The GeoFacial Node, 2018, Public Name and Shame Screen

    Did the idea work, were the audiences faces detected, collected, distributed and used without their permission or control? Were these faces appropriated into the artwork to become an integral part of the work? Did the audience play up to these ideas once they saw themselves ‘on the big screen’? I think the collection of faces and the manner in which the audience presented themselves for ‘documentation’ speaks for itself. Without the audience the work was a static structural object with a soundtrack. Yet even as a static work, the mirrors and the distortions they held, played with the light and individual reflections; triggering that narcissistic awareness of ourselves in the public space. However, once there is a recording and playback device present, the audience up the ante for their own pleasure and those that watch on, in the moment, or later in time. Testing Grounds proved to be a great location to trial this work, to overcome the initial hurdles presented by service requirements and software, and to build my confidence to take this work out of the confines of a ‘safe’ space.

    Another artist working with these ideas of observation is Irish photographer Richard Mosse; his work Incoming (Mosse, 2017) is a 52-minute, 3 channel video work, documenting the plight of refugees following the mass migration and human displacement in Europe - 2014 - 2017. What makes Mosses’ work so captivating is the method in which he chose to capture the refugees. Rather than using standard video technology, Mosse ‘used a military-grade camera designed for battlefield situational awareness and long-range border surveillance in an attempt to confront the viewer with the ways in which our governments represent – and therefore regard – the refugee. We wanted to use the technology against its intended purpose to create an immersive, humanist art form, allowing the viewer to meditate on the profoundly difficult and frequently tragic journeys of refugees’. What this produced was one of the most dramatic video works I have ever encountered. It was alarming, for the maximum functional distance between the camera and the individual human is up to 30km, and disturbing in the way that the human form is anonymous. Not anonymous in the sense that they were unidentified but by the very nature that they were unidentifiable. The captured footage from this military-grade camera was thermal, black and white with the heat being represented by black. This produced what could be best described as an analog camera negative effect, allowing the emotion to be detected in refugee's faces, but not their identity. The GeoFacial Node was conceived originally as a I.D checking station, but in contrast to Mosse's work and the covert equipment he used, this machine was intended from the very start to be obvious in its intentions. The scrolling banner was a play on the 'smart' parking meters but its main point was to advertise the machines intentions, as absurd as they are.

    Several other unexpected observations came out of the installation: firstly - the public name and shame monitor scrolled through images at 20 frames per second, one audience member commented that watching the screen made them consider the frequency that that the human eye scans a crowd for a familiar face. Why did I set the frequency at 20 hertz? Perhaps this was my personal threshold for spotting the image refresh, and in doing so, humanising the process. The speed governments such as China are scanning their crowds at I do not know, but what is obvious it that the scan can be attempting to recognise all faces in the crowd simultaneously. Secondly -The GeoFacial Node acted as a type mapping device. We often think of a map as a means to navigate foreign space, be it a suburb or country. The map talks to ideas of other places. The paper map has been sidelined by the general population for digital alternatives, smart phones and dedicated GPS devices kept close at hand; no longer is the map merely an item to assist you to find your way, the digital alternatives tell you where you are, how to get to your destination and even where you have been. However, I argue that The GeoFacial Node acts as a mapping device by provide a very detailed pixel of a specific location; of a larger image. If this device was deployed on a wide scale rollout, could a map not be generated with the location of entire population at single point of time? 

    It is this unprecedented roll-out of CCTV infrastructure in countries like China, that when integrated with a Social Credit System, provide the authorities with world’s largest living map, mapping the movements of its estimated 1.4+ billion residents. What power does that map bring to those that hold it? I have tried to refrain from making any Orwellian references during this research project as it is a comparison often heard in the media; however, it is hard to overlook the similarities between Orwell’s technological predictions in his novel Nineteen Eighty-Four (Orwell, 1949) and the present day CCTV situation. Orwell wrote about a fictional device called a ‘telescreen’, an all seeing, all hearing, communication device that was installed in public spaces and private homes; monitored and controlled by the authorities. The ‘telescreen’ was also a type of television (smartphone?) that provided entertainment and propaganda; with consequences for those that did not respond accordingly. Thankfully, the ‘telescreen’ is still a fictional device, at least as a singular, stand-alone form. As we embrace technology and willingly invite it into our homes in the form of Google Home and its equivalents, these devices that funnel information down the same networks as our social media feeds and the CCTV footage, one must ask - just how far away is the ‘telescreen’?

    Academics Nicholas Davis and Aleksandar Subic argue(Davis and Subic, 2018) that there is an alternative to Orwell’s future and suggests the solution isn’t so far-fetched. Firstly - we need to take an active role to shape the future - through the design, commercialising, marketing and consumption of this technology, secondly - develop cross- section collaboration between the public and private sectors, and finally, the most relevant to facial recognition software - to ‘tackle the moral component of emerging technologies’. 

    Artists are also challenging these ideas of an Orwellian existence of mass surveillance and observation.  Mexican-Canadian artist Rafael Lorano-Hemmer’s work Zoom Pavilion(Lozano-Hemmer, 2015) tracks the audience as they enter a gallery, singles them out through face tracking and digital zooming and then projects their zoomed face onto the gallery wall, confounding their spatial relationship with predatory technologies of detection and control. London based artist Zach Blas has produced work focusing, not only on the collection of facial biometrics, but delves deeper into how this data is compiled. His work Facial Weaponization Suite is ‘a protest against biometric facial recognition – and the inequalities these technologies propagate’ (Blas, 2011-14); it is a collection of ‘masks’ made up of the faces of individual from groups threatened by this technology, due to race, gender, sexuality or nationality; the mask is then mutated to make it unrecognisable to any biometric technology, allowing the wearer to not be judged. 

    This technology is here to stay and through this work I have proven just how accessible it is to the layman. Obviously, I have an interest in the working of this technology and do honestly believe that there is some good to come out of the use of it. However, the capitalist mentality of our society generally drives advancements in this field that can be sold first and compared to a moral scale later. I do not want to live in a society that has the level of surveillance that China has; I do not want to live my life through social media, but I accept these technologies are here and they will continue to evolve and envelope my life beyond my control. Through conversation and an art practice, I hope that the society I live in can remain aware and subjective of such technologies, keeping them morally and ethically in line.

Simon Reis