IMAGINARIES OF MACHINE VISION. A Short History J E N S S C H R Ö T E R 1. INTRODUCTION Stanley Kubrick’s movie 2001: A Space Odyssey came out in 1968. This famous film has been the subject of a great deal of discussion. An important aspect of the plot is the spaceship crew’s fight with the hyperintelligent computer “HAL”; several times we see HAL’s point of view. This is mostly achieved by first showing one of HAL’s “red eyes” (one of his cameras) and then cutting to the subjective shot, which is marked as the machine’s subjective point of view (or “POV”) by a fish-eye, circular distortion. The film thereby shows us machine vision.1 In 2001, the machine vision shots connote the consciousness of the supercomputer and its active role in the plot. As is well known, HAL becomes paranoid and attempts to kill the crew of the spaceship. How can we analyze this fictional machine vision? What is the role of such a fiction? How was it possible, as early as 1968, to think of such a kind of machine vision? In the second part of this paper, I will discuss some selected positions from the theory of technological imaginaries that might prove helpful for understanding such imaginary technologies. In part three, I will discuss 2001: A Space Odyssey. In part four, I will analyze Westworld (USA 1973, Michael Crichton) before examining, in part five, The Terminator (USA 1984, John Cameron). I will also explain the reasons for these choices. In the conclusion, I will provide a brief analysis of Kazuo Ishiguro’s novel Klara and the Sun (2021). 2. TECHNOLOGY AND IMAGINARIES: THEORIES. The concept of the ‘imaginary’ has a certain role in political philosophy (cf. Taylor 2002).2 As the political scientist Benedict Anderson in particular has demonstrated, the “nation” is an “imagined political community – it is imagined as both inherently limited and sovereign” (Anderson 1983, 6). A similar imaginary component can also be observed in the history of technology. The history of a given technology is also a history of its imaginaries, which means the history of the promises and fears, anticipated futures and expected uses connected to it. Since there has already been 1 See Branigan (1984) on a general theory of subjective shots and especially p. 105 on the “metaphorical POVs” in 2001 and Westworld; see Schröter (2004a) for a philosophical discussion on the fact that films can present subjective experience of others, an access to subjectivity that cannot be experienced in real life. 2 The following is elaborated in much more detail in Ernst/Schröter (2021a). I express my gratitude to Christoph Ernst for writing large parts of that book. NAVIGATIONEN T E C H | D E M O JENS SCHRÖTER a lot of research on this topic, I will limit my discussion to only a few relevant positions. The media studies scholars Simone Natale and Gabriele Balbi differentiate between (1) imaginaries that project future media before they emerge, (2) imaginaries which develop when a medium is real and new, and (3) imaginaries that nostalgically refer to “old” media (cf. Natale/Balbi 2014). The first case is especially relevant to this paper. 2001 imagines a perceiving super computer, a technology that did not exist at the time. In that sense, it represents a form of a “media prophecy” (ibid., 205–207). Media prophecies have existed at least since the nineteenth century, especially in science fiction. Science fiction is a discourse that – as Arthur C. Clarke (1962) noted – reaches out imaginatively into the impossible in order to determine what is possible. One example of this is the relationship between the history of computing and science fiction. For the computer scientists David L. Ferro and the historian Eric G. Swedin (2011), science fiction functions as a link between the public and the development community. Indeed, in 2001, some members of the community developing artificial intelligence were scientific consultants (cf. Kirby 2003). On the one hand, science fiction helps us imagine computing’s possible consequences for society; on the other hand, it makes explicit the “needs,” that is, the wishes and desires, that drive development in society (cf. Ferro/Swedin 2001). In particular, the concept of imaginaries contains an understanding of the instrumentalization of science fiction as part of the “strategic” planning of possible futures. Today, science fiction is very consciously employed in technology development and product design through “science fiction prototyping” (Johnson 2011). It is therefore interesting to observe science fiction as one site of imaginaries of machine vision, although it is an open question at this point if HAL in 2001 can be called a case of science fiction prototyping. The goal of science and technology studies (STS) is to study the process of the social production of science and technology (cf. Sismondo 2010). The role of imaginaries has been a matter of intense discussion in this research tradition in recent years (Sneath, Holbraad, and Pedersen 2009; McNeil et al. 2017). The term “sociotechnical imaginaries” developed by the historians of science Sheila Jasanoff and Sang-Hyun Kim (2015) is a case in point. Jasanoff (2015) regards sociotechnical imaginaries as “collectively held, institutionally stabilized, and publicly performed visions of desirable futures, animated by shared understandings of forms of social life and social order attainable through, and supportive of, advances in science and technology.” The concept of sociotechnical imaginaries illuminates historical changes in ideas about the interplay of sociality and technology. If one compares this very broadly conceived concept to the concept of the Leitbild (guiding image), which is common in the German-speaking sociology of technology, the similarities are striking. Katharina Giesel writes: “Leitbilder bring together socially separated NAVIGATIONEN 94 T E C H | D E M O IMAGINARIES OF MACHINE VISION (mental or verbalized) ideas of a desired or desirable and in principle achievable future that becomes reality though corresponding action” (Giesel 2007, 245, my translation). In both cases, it is a matter of analyzing the collective imagination about a desirable future for science and technology. The concept of “sociotechnical imaginaries” emphasizes the aspects of institutional safeguarding and of public performance (mise-en-scène, staging, presentation). Both concepts focus on the mediating function of imaginaries for coordinating in the development of new media and the social acceptance of these technologies. Therefore, one question I wish to address in my analyses below is: To which socio-technical imaginaries or Leitbilder are fictional representations of machine vision connected?3 Since there is an emphasis on the public performance of socio-technical imaginaries or Leitbilder (literally: guiding images) in STS, this research specifically includes approaches to analyzing the media required for presenting imaginaries, such as theatre-like practices, e. g. stages, and other forms of “negotiation arenas” (Schulz-Schaeffer and Meister 2009) in which, for example, prototypes are choreographed and displayed (cf. Ernst and Schröter 2021b). The STS scholar Wally Smith has presented an approach to analyzing such presentations under the title “theatre of use” (Smith 2009, 449). In the IT industry, demonstrations serve to show hardware and software in action. In front of a heterogeneous circle of addressees, developers present the evidence that the technology works and is useful. The historical model for this is public demonstrations of scientific experiments (cf. ibid., 451–457). One can observe, according to Smith, a difference between the way significance is attributed to an experiment in the everyday practice of science and the significance an experiment takes on when it is publicly demonstrated. Using Erving Goffman’s (1974) frame analysis, now well established in sociology, Smith understands this as a “reframing.” Whereas a laboratory experiment is about representing nature, public demonstrations are about representing the possibilities of the technical apparatus and of the experiment itself (cf. Smith 2009, 456). The focus is not on what could be found out about nature in the future but on the future possibilities of technological feasibility (cf. ibid., 453). One example is the gesture control of the interface in Steven Spielberg’s science fiction thriller Minority Report (Spielberg 2002). The interface designer John Underkoffler had already developed experimental prototypes of this technology in reality. In Spielberg’s film, however, the interface is presented as a completely functional technology with idealized scenarios of possible use. STS scholar David Kirby refers to such phenomena as “diegetic prototypes.” (Kirby 2011, 193–218). Diegetic prototypes are objects that are fully functional in 3 At the University of Bergen in Norway, there is an EU research project entitled “Machine Vision in Everyday Life: Playful Interactions with Visual Technologies in Digital Art, Games, Narratives and Social Media”, https://www.uib.no/en/machinevision, accessed June 3, 2022. They have built up an interesting database: https://machine-vision.no, accessed June 3, 2022. Their research interests and publications (that can be found on the website) do not overlap with this paper. NAVIGATIONEN T E C H | D E M O 95 JENS SCHRÖTER the diegetic world constructed by the narration. In the context of imaginaries, they mobilize “public support for potential or emerging technologies by establishing the need, benevolence, and viability of these technologies” (ibid., 18). We can summarize Kirby’s arguments regarding the representation of imaginary technologies in a film in the following four points: 1. Performative artifacts: The objects demonstrate the possibilities of an emerging technology; 2. Social contextualization: The objects are situated by the narrative in a social context; for instance, they are associated with a community of users; 3. Normalized approach: The objects are presented as completely normal objects and their use as completely ‘natural;’ 4. Real need: The technology is framed in such a way that the film may trigger a real need for these objects. The popular science fiction film is a widespread and influential example of such presentations. Science fiction films are conveyors of technology and media for reflecting on the technological change that is taking place. What “theatres of use” and “diegetic prototypes” can we find in representations of machine vision? In the following sections I will to return the fictional representation of machine vision in 2001, but will also consider other examples of an imaginary machine vision in science fiction and analyze them in their respective contexts. How is machine vision represented? Which socio-technical imaginaries are connected to it? Can we find Kirby’s different aspects in the filmic representations? 3. 1968: 2001: A SPACE ODYSSEY [Fig. 1: HAL’s gaze] NAVIGATIONEN 96 T E C H | D E M O IMAGINARIES OF MACHINE VISION In 1960, one of the earliest papers on machine vision was published: “Pattern Recognition by Machine” by Oliver G. Selfridge and Ulric Neisser. It appeared in Scientific American – in a journal, then, that popularized current scientific knowledge. The paper opens with the following passage: Can a machine think? The answer to this old chestnut is certainly yes: Computers have been made to play chess and checkers, to prove theorems, to solve intricate problems of strategy. Yet the intelligence implied by such activities has an elusive, unnatural quality. It is not based on any orderly development of cognitive skills. In particular, the machines are not well equipped to select from their environment the things, or the relations, they are going to think about (Selfridge and Neisser 1960, 60). In other words, the machine needs a capability for perception or, as the authors put it, pattern recognition. Although the text focuses on the recognition of printed and handwritten characters, this is clearly a case of machine vision. Interestingly, they start with the question of “thinking machines”, which had been a somewhat fashionable topic in computer science since the Dartmouth Conference of 1956. Through popularizations such as Selfridge and Neisser’s text, “artificial intelligence” had become a well-known theme (an “old chestnut”) by the end of the 1960s. Indeed, one of the protagonists of the Dartmouth Conference, the computer scientist Marvin Minsky, was a scientific consultant to 2001. He is especially known for advocating a modular concept of mind in which different specialized “modules” interact to produce “intelligence” (cf. Minsky 1988).4 One such module is perception. The staging of the supercomputer HAL as possessing the ability to understand and produce spoken language, to play games, to plan, and to see shows his intelligence to be of such a modular kind.5 We can state that machine vision appears in 2001 in the context of a movie that tries to be scientifically accurate about (a certain modular conception of) AI.6 HAL’s POV is staged as a ‘normal’ view (it looks like the POVs of the human actors, as compared to other more technologized POVs I will discuss below), except for the fact that it has a circular form that mimics the circular form of his red “eyes”, which are dispersed all over the spaceship (Fig. 1). At the time, there were certainly not any digital tricks available for staging his POV (as compared to the machinic POV in Westworld 1973, see below), but it would of course have been possible to 4 This book was published twenty years after 2001. 5 See Mateas (2006, 107): “Marvin Minsky, one of the founders of AI, served as a technical consultant for the film: doubtless his contribution helped to establish the strong resonance between the depiction of HAL and subfields within AI, including language, commonsense reasoning, computer vision [!], game playing and planning, and problem solving.” See also Stork (1997). 6 See Rosenfeld (1997) for a detailed analysis of the representation of machine vision in 2001 from the standpoint of a – leading – practitioner in the field. NAVIGATIONEN T E C H | D E M O 97 JENS SCHRÖTER defamiliarize HAL’s view in some other ways. Staging it as very similar to the point- of-view shots of the human protagonists would have been a very plausible means to show HAL’s intelligence as comparable to that of humans – and it also implies self-awareness on the part of the super computer. I now wish to systematically apply Kirby’s methodological steps for analyzing diegetic prototypes: (1) Performative artifacts: 2001 demonstrates the possibilities of an emerging technology, AI. But as Kirby underlines, to work as a diegetic prototype, a technology has to be shown to work flawlessly– and that is of course what the film does not do. As is well known, HAL becomes paranoid and kills the crew (with the exception of Dave, who, in the end, shuts HAL down). This is first and foremost for dramaturgical reasons, of course, because an antagonist is needed (as in Westworld, see below). However, to stay with Kirby’s reasoning, it demonstrates that the film does not intend to arouse interest in and acceptance of AI. AI is simply presupposed to be a real technology in a distant (or not so distant: 2001) future and to have ambiguous implications. Machine vision in this context can be read as a potentially dangerous technology of surveillance – think of the scene in which HAL reads the astronauts’ lips to thwart their plans. (2) Social contextualization: In 2001, the mission has a scientific character, but it seems also to have a hidden military background, since the astronauts were not fully informed about its real purpose. This points to real developments: as Yarden Katz points out, there was a memo at the “MIT Artificial Intelligence Lab, one of the first academic AI laboratories (cofounded by Minsky), written in July 1966. The memo, authored by Seymour Papert, describes plans to build – in the scope of one summer – an artificial vision system that could recognize objects.” Furthermore, he adds: “Practitioners had some reason for optimism: the term ‘Artificial Intelligence’ charmed the Pentagon’s elites, and they poured money into the field” (Katz 2020, 24). Finally, Katz notes that “AI’s subfields were from the start organized around a militaristic frame: vision research to detect ‘enemy’ ships and spot resources of interest from satellite images […]” (ibid., 35). These real developments informed the film, which hints at a special community of users in high science and the military, and locates machine vision in this context – and HAL (be he paranoid or not) indeed tries to kill the astronauts in a ruthless, military manner to secure the mission. (3) Normalized approach: For the astronauts it is quite normal to communicate with HAL and to accept that he has a kind of vision. In one scene, HAL asks astronaut Dave Bowman to show him a drawing he has done. Dave complies and HAL comments on the drawing.7 Machine vision is normalized – at least for special users like astronauts. Nonetheless, this technology is not safe: The gaze of the AI turns out to be a terrorizing instrument of control. The AI indeed ends up killing most of the crew, an outcome that foreshadows contemporary 7 The scene can be seen here: https://www.youtube.com/watch?v=_jeJk63XYPg, accessed June 3, 2022. NAVIGATIONEN 98 T E C H | D E M O IMAGINARIES OF MACHINE VISION discussions about the danger of autonomous weapon systems that have capabilities of perception and select targets to destroy them. In not only normalizing but subsequently showing the dangers of AI, the film ultimately departs from Kirby’s typology, since, for Kirby, “movies can show audiences how a technology works, why it is safe and why they need it” (Kirby 2010, 45). (4) Real need: Since HAL and its capabilities are located in a special, high science, space travel, and at least implicitly military context, it does not correspond or coincide with real-life needs, such as those of the audience, for instance. Machine vision is presented/interpreted as removed from ordinary life. As Michael Mateas notes, the figuration of HAL turned out to be a quite powerful socio-technical imaginary: “As a representation, HAL, and the role he plays within 2001, both captures preexisting intellectual currents that were already operating with the field of AI and serves as an influential touchstone, which has had a profound impact on individual AI practitioners and on the aspirations of the field” (Mateas 2006, 105). The idea of machine vision was connected, at an early juncture, with the ambiguous idea – unrealized to this day – of a general artificial intelligence. 4. 1973: WESTWORLD [Fig. 2: Subjective of the Gunslinger robot] In 1973 Westworld was released, a science fiction movie directed by Michael Crichton. It presents a different scenario than 2001. The “Westworld” is a kind of high-tech amusement park in which guests can play, often transgressively, with lifelike androids in different historical settings. Then a malfunction happens: The security measures fail and the androids become a serious threat to the guests. The antagonist, an artificial gunslinger played by Yul Brynner, kills one guest, chases another, but can be defeated at the end. The historical background informing the film is similar to that of 2001: “Artificial intelligence” was a widespread – albeit controversial – topic in 1973. The idea of self-acting robots was not completely NAVIGATIONEN T E C H | D E M O 99 JENS SCHRÖTER alien. But Westworld adds some new twists to the picture. To illustrate this, I turn, again, to Kirby’s scheme: (1) Performative artifacts: In Westworld, machine vision is connected not to an all-seeing AI but to an always situated robot that is, on the one hand, very limited compared to HAL. This robot is only a gunslinger and completely reduced to that task – in that sense he is an imaginary representation of a “narrow AI” compared to HAL, which is an imaginary representation of a “general AI” (cf. Dyer-Witheford, Kjøsen and Steinford 2019, 10–15). On the other hand, the machine is able to move and, therefore, to chase the protagonist, showing working robotics at play. In this context we see the POV of the robot, its machine vision, several times (Fig. 2).8 While HAL is an image of a symbolic AI, the gunslinger in Westworld is an embodied AI. Machine vision is constructed here as robots’ means of orienting themselves in space. Interestingly, Westworld is now considered to be the first film in which digitally processed images were used to present the robot’s POV. Computer graphics pioneer John Whitney Jr. writes: When I met Brent Sellstrom, the post-production supervisor of Westworld, he had a problem: to find a technique to represent, on film, the point of view of a machine. The script called for the audience to see the world as a robot gunfighter, played by Yul Brynner, saw it. This ‘Robot POV’ was supposed to consist of a series of animated colored rectangles. It could not be done by any known special-effects technique. Something new was required (Whitney Jr. 2020).9 Yet, he adds: It occurred to me that the scanning digitizing methods employed by NASA’s Jet Propulsion Laboratory on their Mariner Mars flybys could be used here. Basically, in this system, an image is broken down into a series of points, and the gray-scale value for each point determined. A numerical value can then be assigned to each point, and a new image reconstituted electronically. Similar techniques have been devised in computer science to enable computers to ‘read’ handwriting, X-rays, seismic data, and so on. It is the kind of technology that allows a computer to tell the difference between a ‘P’ and an ‘R’ (ibid.).10 8 See https://www.youtube.com/watch?v=nILKFlpOZi0, accessed June 3, 2022 for a compilation of machine vision scenes from Westworld. 9 Whitney Jr. (together with Saul Bass) also created the very famous graphic title sequence of Vertigo (Alfred Hitchcock, USA 1958), the first ever computer graphic sequence in cinema (cf. McCormack 2013). 10 The details of the cumbersome process are provided in detail in that article. See also ASC Staff (2020). NAVIGATIONEN 100 T E C H | D E M O IMAGINARIES OF MACHINE VISION Ultimately, Whitney had to use another approach than that deployed by NASA, for simple reasons of cost. Yet the images representing the machine vision in Westworld refer to the history of the digitization of images, which would turn out to be very important not only for science, space travel etc. (cf. Schröter 2004b) but also, as he explicitly states, for the history of pattern recognition, that is: machine vision itself. The POV shots of the robot not only represent the vision of a machine but are actually made with technologies that finally led to machine vision. (2) Social contextualization: Contrary to 2001, the plot of Westworld shows the advanced technology of machine vision not in space travel and in a scientific and/or military context, but in a future form of entertainment. In that sense, it connects machine vision with the not unusual transfer of technologies from the military to the civil realm. Indeed, Whitney Jr. was known for “creating animation with military-surplus analog electronics and motor assemblies” (Price 2013).11 On the level of content, as on that of material production (at least concerning Whitney’s earlier work, before he did special effects), the film presents the transfer of military technological inventions to civil entertainment. (3) Normalized approach: Westworld describes a future in which it will be normal (at least if you have enough money) to interact with robots that can see. The gunslinger is a slave-like robot designed for a very specific task that then breaks out of his assigned role, not in an act of conscious rebellion, however, but as a result of technical failure. Again, advanced technology is imagined as a potential threat. (4) Real need: In Westworld, the socio-technical imaginary of future entertainment is modeled after Disneyworld: going to a special place, a theme park, and para-socially interacting with fictional, but tangible and in that sense virtual characters in fictional scenarios. The real need for an ever-expanding entertainment industry foreshadows its own future. Such robot parks are still far from becoming real, even if a contemporary television series directly based on Westworld refreshes this imaginary.12 Perhaps this will never become a real practice, yet narrow AIs – like the robot in Westworld – are in fact already widespread today. While robots have mostly non-anthropomorphic forms and are used in industry, simple AI systems possessing the ability to recognize visual and aural patterns are to be found in Alexa or Siri or in the face recognition systems of modern iPhones. Westworld introduced digitally processed images and their digital look to audiences and links this electronic POV to a robotlike, narrow AI. The rhetoric of such technological-looking images is well established today. Machine vision appeared/emerged in Westworld in a context of entertainment, mirroring the then slowly nascent transfer of computer technologies, originally developed by high science and the military, into entertainment and, finally, everyday life. 11 See also McCormack (2013). 12 Westworld (Nolan and Joy 2016). NAVIGATIONEN T E C H | D E M O 101 JENS SCHRÖTER 5. 1984: THE TERMINATOR [Fig. 3: Subjective of the Terminator] This dissemination of computing began in the late 1970s with the creation/foundation of companies such as Microsoft and Apple. One popular computer was the Apple II, released in April 1977. In 1984, The Terminator (directed by James Cameron) opened in the US cinemas. It told the story of a robot, the “terminator”, who is sent from the future to kill the mother (Sarah Connor, played by Linda Hamilton) of the future leader of the resistance against the tyranny of artificial intelligence in that same future. A human fighter, played by Michael Biehn, is sent back too and tries to protect the mother. He then falls in love with her, they have sex – and become the parents of the coming leader. What is the connection to the diffusion of early home and personal computing?13 The subjective POV of the robot from the future is, different from the POVs in 2001 and Westworld, overlayed with diagrams and snippets of computer code to make it look more robot-like (Fig. 3). The code snippets are from 6502 assembler code, which was the exact code used in the Apple II.14 The snippets were taken from the home and personal computer hobbyist journal Nibbles.15 The 6502 processor and its close relative, the 6510, were also used in 13 On the history of early home and personal computing, see Haddon (1988, especially pp. 23–24 for the Apple II). 14 See a compilation of POV scenes from all the Terminator films: https://www.youtube.com/watch?v=CZlEiD6Nmnc, accessed June 3, 2022. 15 E. g. Manly (1984, 52); Peterson (1984, 40); Efflandt (1984, 89). Since the last two examples are from the September issue of the journal Nibble and since The Terminator was released in the US on October 26, 1984, it seems that the POV shots with the snippets of the code were filmed very late and shortly before release. NAVIGATIONEN 102 T E C H | D E M O IMAGINARIES OF MACHINE VISION other popular computers of the time, such as the Commodore VIC 20 and the Commodore C64 (cf. Bagnall 2011). We again use Kirby’s scheme: (1) Performative artifacts: As in Westworld, we are shown a robot which has astounding capabilities but is more limited than HAL. He has only one purpose: killing Sarah Connor. What is more important is that the machine vision of the terminator shows the visual field overlayed with information. This prefigures the now widespread idea of “augmented reality” (AR) (cf. Schröter 2011). Viewed from a contemporary perspective, the robot’s POV displays the enhanced and augmented POV of today’s computer users avant la lettre. It shows the successful performance of an AR system. Moreover, since the Apple II had been available since 1977, and insofar as many users would have known the 6502 code (although most of them presumably programmed in easier higher programming languages), the POV also references another, well-performing, technology: home and personal computers of the time. (2) Social contextualization: Machine vision is shown in a warlike and therefore military context. The robot from the future means the threat of hyperintelligent AI exterminating mankind. In that context, machine vision is, somewhat similar to Westworld, but on a much larger scale, conceptualized as a predatory and even genocidal gaze – a dystopian scenario that is still influential today, for instance when world-famous physicist Stephen Hawking warns against AI (cf. Cellan-Jones 2014).16 (3) Normalized approach: The scenario is obviously different from 2001 and Westworld. The technology is not normal, but comes as a deadly threat from a distant future. However, the film does make it seem to be normal for advanced AI to use killer robots. (4) Real need: Although it may reasonably seem like there is no meaningful need for dangerous killer robots, there is actually a lot of research on autonomous weapon systems conducted by the military. Terminator, in that sense, can be seen as metaphorically representing the automatization of warfare and the role of machine vision as a technology for identifying targets (cf. Virilio 1994; Queisner 2017). Furthermore, the imaginary technology presented in Terminator is removed from Kirby’s notion of the “diegetic prototype”: Diegetic prototypes differ substantially from what I term ‘speculative scenarios’ in movies […]. Speculative scenarios represent highly implausible and impractical situations and technologies that film-makers and science consultants imbue with a sheen of plausibility, so that they look possible within a film’s narrative. They make these technologies look plausible, knowing that they are impossible to achieve in real life. (Kirby 2010, 46) 16 On other dystopian AI scenarios, see Dyer-Witheford, Kjøsen and Steinford (2019, ch. 3). NAVIGATIONEN T E C H | D E M O 103 JENS SCHRÖTER Although Terminator surely has no direct influence on the socio-technical imaginaries or Leitbilder of technological development (compared to HAL from 2001), it was influential in an aesthetic way. With its representation of the machine POV as images overlayed with diagrams and code, Terminator, just like Westworld’s pixilated and ‘technical look,’ has shaped the aesthetic of imaginary machine vision to this day. The sequels to Terminator contain machine POVs that further modify and extend the aesthetics developed in the initial movie. Interestingly, the idea of integrating information snippets in the machine vision was already present in the production of Westworld: Crichton was concerned from the outset with how to give a distinct look to the gunslinger’s point of view when the audience saw events through its eyes. Five years earlier, Stanley Kubrick had used a wide- angle lens to show the perceptions of [HAL] 9000, the troubled computer of ‘2001: A Space Odyssey.’ Crichton, however, wanted his villain’s perspective to look like that of an electronic machine. His script described it as a bizarre, ‘computerized image of the world’ with ‘flashed-up calculated figures’ (Price 2013). 6. A CONCLUSION WITH KAZUO ISHIGURO’S KLARA AND THE SUN What the three films analyzed have in common is that they have shaped the imaginary of machine vision. All subsequent representations in films and television basically resemble their aesthetic and thematic construction of the POV of machines. Through numerous films and television series, the idea of a machine POV (and the associated idea of machine consciousness) has become well established in popular culture and part of our socio-technical imaginaries and media prophecies. The idea of a machine POV and the associated idea of machine consciousness is so well known that today it can even be found in highbrow literature, normally far removed from science fiction. A remarkable example is the novel Klara and the Sun (2021) by Nobel Prize (2017) winner Kazuo Ishiguro. The story is narrated from the point of view of a so-called “artificial friend” (AF), a highly developed robotic, female-connotated AI system named Klara. She is a kind of robot puppet that can be bought as a companion for children. The machine is thus associated with “female” values such as care – in contrast to the “male-connotated” computer systems discussed above.17 Klara tries to understand the world in her own way. The details of the wonderful and deeply moving plot need not be discussed here. Suffice it to emphasize, first, that AI is represented not according to today’s stereotyped representations as either a slave or a threat (except for the fact that many people in the story world are “post-employed” due to technical progress18), 17 This embeds Klara in a history of female-connotated robots in film, e.g. Her (USA 2013, Spike Jonze). 18 On AI and labor, see Dyer-Witheford, Kjøsen and Steinford (2019, ch. 1 and 2). NAVIGATIONEN 104 T E C H | D E M O IMAGINARIES OF MACHINE VISION but as a loving friend and companion. Second, since the book is centered around Klara’s subjective experience, it is also centered around her visual (and aural) perception. At the very beginning, where she is presented in a store window and object of the customers’ gaze, we read: “I should confess here that for me, there’d always been another reason for wanting to be in the window […] Unlike most AFs […] I’d always longed to see more of the outside – and to see it in all its detail” (Ishiguro 2021, 6).19 Many passages in the narration emphasize the way she visually perceives the world and it becomes obvious that Klara’s field of vision is technologically pixilated in a way that resembles the POVs in Westworld: “The mother leaned closer over the tabletop and her eyes narrowed till her face filled eight boxes, leaving only the peripheral boxes for the waterfall, and for a moment it felt to me her expression varied between one box and the next” (ibid., 104). Her field of vision is somehow structured in “boxes.” Later on, Klara has some problems with her eyesight: I’d thought that once I was no longer observing them through glass, the theater people would become more distinct. But now I was in their midst, their figures became more simplified, as if constructed out of cones and cylinders made from smooth card. Their clothes, for instance, were devoid of the usual creases and folds, and even their faces under the streetlight appeared to have been created by cleverly placing flat surfaces into complex arrangements to create a sense of contouring (ibid., 235). If the idea that machines can see, like the recent iPhones that can recognize our faces, has become a well-established topic, it is because more than fifty years of socio-technical imaginaries of machine vision have accustomed us to the idea that vision is no longer the privilege of higher biological life forms alone. REFERENCES Anderson, Benedict. 1983. Imagined Communities: Reflections on the Origin and Spread of Nationalism. London: Verso. ASC Staff. 2020. “Behind the Scenes on Westworld: AC Talks to Writer-Director Michael Crichton.” American Cinematographer, March 12, 2020, https://ascmag.com/articles/behind-the-scenes-on-westworld. Bagnall, Brian. 2011. Volkscomputer: die Geschichte von Pet und VC-20, C64 und Amiga. Aufstieg und Fall des Computer-Pioniers Commodore. Utiing: Gameplan. 19 See also Ishiguro (2021), where one character says to Klara: “You’re an intelligent AF. Maybe you can see things the rest of us can’t” (108). NAVIGATIONEN T E C H | D E M O 105 JENS SCHRÖTER Branigan, Edward. 1984. Point of View in the Cinema: A Theory of Narration and Subjectivity in Classical Film. Berlin: Mouton Publishers. Cameron, James, director. 1984. The Terminator. Cinema ’84. 1 hr., 47 min. Cellan-Jones, Rory. 2014. “Stephen Hawking Warns Artificial Intelligence Could End Mankind.” December 2, 2014, https://www.bbc.com/news/technology- 30290540. Clarke, Arthur C. 1962. Profiles of the Future: An Inquiry into the Limits of the Possible. New York: Harper & Row. Crichton, Michael, director. 1973. Westworld. Metro-Goldwyn-Mayer. 1 hr., 28 min. Dyer-Witheford, Nick, Atle Mikkola Kjøsen, and James Steinford. 2019. Inhuman Power. Artificial Intelligence and the Future of Capitalism. London: Pluto. Efflandt, David. 1984. “Solid State Slide Show.” Nibble 5 (9): 87–90. Ernst, Christoph, and Jens Schröter. 2021a. Media Futures. Theory and Aesthetics. Cham: Palgrave Macmillan. Ernst, Christoph, and Jens Schröter. 2021b. “Die Zukunft vorstellen. Technologie- Demonstrationen in der Geschichte digitaler Medien.” Technikgeschichte 88 (1): 79–105. Ferro, David L., and Eric G. Swedin, eds. 2001. Science Fiction and Computing: Essays on Interlinked Domains. Jefferson, NC: McFarland. Giesel, Katharina D. 2007. Leitbilder in den Sozialwissenschaften. Begriffe, Theorien und Forschungskonzepte. Wiesbaden: Springer VS. Goffman, Erving. 1974. Frame Analysis: An Essay on the Organization of Experience. New York: Harper & Row. Haddon, Leslie. 1988. “The Home Computer: The Making of a Consumer Electronic.” Science and Culture 1 (2): 7–51. Ishiguro, Kazuo. 2021. Klara and the Sun. London: Faber and Faber. Jasanoff, Sheila. 2015. “Future Imperfect: Science, Technology, and the Imagination of Modernity.” In Dreamscapes of Modernity: Sociotechnical Imaginaries and the Fabrication of Power, edited by Sheila Jasanoff and Sang-Hyun Kim, 1–33. Chicago: Univ. of Chicago Press. Jasanoff, Sheila, and Sang-Hyun Kim. 2015. Dreamscapes of Modernity: Sociotechnical Imaginaries and the Fabrication of Power. Chicago: Univ. of Chicago Press. Johnson, Brian David. 2011. Science Fiction Prototyping. Designing the Future with Science Fiction. San Rafael, CA: Morgan & Claypool. Katz, Yarden. 2020. Artificial Whiteness. Politics and Ideology in Artificial Intelligence. New York: Columbia University Press. NAVIGATIONEN 106 T E C H | D E M O IMAGINARIES OF MACHINE VISION Kirby, David A. 2003. “Science Consultants, Fictional Films, and Scientific Practice.” Social Studies of Science 33 (2): 231–268. https://doi. org/10.1177/ 03063127030332015. Kirby, David A. 2010. “The Future Is Now: Diegetic Prototypes and the Role of Popular Films in Generating Real-World Technological Development.” Social Studies of Science 40 (1): 41–70. https://doi. org/10.1177/0306312709338325. Kirby, David A. 2011. Lab Coats in Hollywood: Science, Scientists, and Cinema. Cambridge, MA: MIT Press. Kubrick, Stanley, director. 1968. 2001: A Space Odyssey. Metro-Goldwyn-Mayer. 2 hr., 29 min. Manly, Ken. 1984. “Mon.E. A 128 K Memory Examination Tool.” Nibble 5 (8): 45- 61. Mateas, Michael. 2006. “Reading Hal: Representation and Artificial Intelligence.” In Stanley Kubrick’s 2001: A Space Odyssey: New Essays, edited by Robert Kolker, 105–125. Oxford University Press. McCormack, Tom. 2013. “Did ‘Vertigo'’ Introduce Computer Graphics to Cinema?” Rhizome, May 9, 2013, https://rhizome.org/editorial/2013/may /9/did-vertigo-introduce-computer-graphics-cinema/. McNeil, Maureen, Michael Arribas-Ayllon, Joan Haran, Adrian Mackenzie, and Richard Tutton. 2017. “Conceptualizing Imaginaries of Science, Technology, and Society.” In The Handbook of Science and Technology Studies, edited by Ulrike Felt, Rayvon Fouché, Clark A. Miller, and Laurel Laurel Smith-Doerr, 435–463. Cambridge, MA: MIT Press. Minsky, Marvin. 1988. The Society of Mind. New York: Simon & Schuster. Natale, Simone, and Gabriele Balbi. 2014. “Media and the Imaginary in History.” Media History 20 (2): 203–218. Nolan, Jonathan, and Lisa Joy, creators. 2016. Westworld. HBO. Peterson, Douglas C. 1984. “Ram Disk 64” Nibble 5 (9): 33–42. Price, David A. 2013. “How Michael Crichton’s ‘Westworld’ Pioneered Modern Special Effects.” The New Yorker, May 14, 2013, https://www.newyorker .com/tech/annals-of-technology/how-michael-crichtons-westworld- pioneered-modern-special-effects. Queisner, Moritz. 2017. “‘Looking Through a Soda Straw’: Mediated Vision in Remote Warfare.” Politik 20 (1): 1–18. Rosenfeld, Azriel. 1997. “Eyes for Computers. How HAL Could ‘See’.” In HAL’s Legacy. 2001’s Computer as Dream and Reality, edited by David G. Stork, 211–236. Cambridge/MA and London: MIT Press. Schröter, Jens. 2004a. “Alterität und Medialität. Ein Versuch zwischen transzendentaler Phänomenologie und Medientheorie.” Navigationen 4 (1/2): 11–26. NAVIGATIONEN T E C H | D E M O 107 JENS SCHRÖTER Schröter, Jens. 2004b. “Das Ende der Welt. Analoge und Digitale Bilder – mehr oder weniger Realität?” In Analog/Digital – Opposition oder Kontinuum? Beiträge zu Theorie und Geschichte einer Unterscheidung, edited by Jens Schröter and Alexander Böhnke, 335–354. Bielefeld: Transcript. Schröter, Jens. 2011. “Echtzeit und Echtraum. Zur Medialität und Ästhetik von ‚Augmented Reality‘-Applikationen.” Augenblick – Marburger Hefte zur Medienwissenschaft (3): 104–120. Schulz-Schaeffer, Ingo and Martin Meister. 2009. “Prototype Scenarios as Negotiation Arenas Between the Present and Imagined Futures. Representation and Negotiation Power in Constructing New Socio-Technical Configurations.” In Socio-Technical Futures Shaping the Present. Empirical Examples and Analytical Challenges, edited by Andreas Lösch, Armin Grunwald, Martin Meister and Ingo Schulz-Schaeffer, 37–65. Wiesbaden: Springer VS. Selfridge, Oliver G. and Ulric Neisser. 1960. “Pattern Recognition by Machine.” Scientific American 203 (2): 60–69. Sismondo, Sergio. 2010. An Introduction to Science and Technology Studies. Malden: Wiley-Blackwell. Smith, Wally. 2009. “Theatre of Use: A Frame Analysis of Information Technology Demonstrations.” Social Studies of Science 39 (3): 449–480. Sneath, David, Martin Holbraad, and Morten Axel Pedersen. 2009. “Technologies of the Imagination: An Introduction.” Ethos 74 (1): 5–30. https:// doi.org/10.1080/00141840902751147. Spielberg, Steven, director. 2002. Minority Report. 20th Century Fox. 2hr., 25 min. Stork, David G. 1997. “Scientist on the Set: An Interview with Marvin Minsky.” In HAL’s Legacy. 2001’s Computer as Dream and Reality, edited by David G. Stork, 15–32. Cambridge/MA and London: MIT Press. Taylor, Charles. 2002. “Modern Social Imaginaries.” Public Culture 14 (1): 91–124. Virilio, Paul. 1994. The Vision Machine. Perspectives. London: BFI. Whitney Jr., John. 2020. “Creating the Special Effects for Westworld. The Inventive Approach to Rendering a Killer Robot’s Electronic POV for this Imaginative Sci-fi Tale.” American Cinematographer, March 15, 2020, https://ascmag.com/articles/creating-the-special-effects-for-westworld. IMAGE INDEX Fig. 1: Taken from the Film 2001: A Space Odyssey (Kubrick 1968). Fig. 2: Taken from the Film Westworld (Crichton 1973). Fig. 3: Taken from the Film The Terminator (Cameron 1984). NAVIGATIONEN 108 T E C H | D E M O