Algorithmic tyranny: Psycho-Pass,science fiction and the criminological imagination

Introduction

In ‘Social science fiction’, Isaac Asimov (1953), one of the genre’s most celebrated and prolific authors, delineates three categories into which all science fiction stories may be placed. In the gadget plot, the story’s focus remains on the new technology itself. In the adventure plot, the gadget serves as a dramatic prop that either helps, hinders or causes the central problem characters must overcome. Finally, in the social plot, the story focuses on the effect a new gadget or technology has on the everyday lives of characters, and ‘the impact of scientific advance upon human beings’ (Miller, 1977: 14). Social science fiction is everywhere. We see it in Orwell’s celebrated 1984, and Huxley’s Brave New World. We can find it in Netflix’s Black Mirror, and in an array of contemporary TV shows, films and novels. And increasingly, we can find it in the work of social scientists themselves, where it not only offers concepts, but also speculative visions of futures and social orders moulded by new technologies (Burrows, 1997; Frase, 2016; Jameson, 2005).

Social science fiction has much to offer criminologists. Through offering people-centric visions of technologically-driven futures, the genre has sparked technologists’ imaginations, offered policy makers cautionary tales and provided perhaps the central wellspring of narratives about technological futures. We have seen on numerous occasions already how social science fiction has been assimilated into academic criminology. Orwell’s totalitarian dystopia of 1984 gifted criminologists and sociologists alike with the notion of Big Brother-style surveillance (see Lyon, 2003; Marks, 2005). Phillip K Dick’s work has furnished criminologists with the concept of pre-crime (McCulloch and Pickering, 2009; McCulloch and Wilson, 2016; Zedner, 2007a). Even the term ‘cyberspace’, much used within cyber-criminology (see Jaishankar, 2007; Ngo and Paternoster, 2011; Wall, 2007), has its origins in Gibson’s seminal novel Neuromancer.

The last decade has witnessed a surge in studies of ‘popular criminology’: the theories of crime and criminal justice advanced within cinema, literature and other forms of popular culture (Kohm, 2017; Rafter, 2007). Yet whilst criminal justice has been a longstanding theme within science fiction and literature studies (see Olander and Greenberg, 1977), few criminologists have examined the diverse literary and cinematic genre of science fiction (see Brown, 2010; Cohen, 1985; Wall, 2008). Following Cohen’s (1985) lead, in this article I want to suggest that science fiction’s ability to help us think ‘longer term about technology’ (Miller and Bennett, 2008: 598) can offer another important resource for criminology: a means to unpack ‘visions of social control’ underpinned by technological advancements. Science fiction, I argue, offers a springboard for imagining, interrogating and sensitizing us to the technological unconscious of criminal justice and social control – the opaque workings of technologies employed within the criminal justice sector that are rarely understood by the individuals who employ them to detect, predict, map and respond to crime.

This article makes a case for the value of science fiction to criminologists through examining the celebrated and popular cyber-punk anime series Psycho-Pass. Written by Gen Urobuchi and directed by Naoyoshi Shiotani, Psycho-Pass, which ran between 2012 and 2014, ¹ depicts a dystopian Japan governed by an advanced pre-crime supercomputer. As a work of both social science fiction and hard science fiction – science fiction that attempts to remain faithful to the laws of nature and ‘committed to avoiding scientific errors in stories’ (Westfahl, 1993: 162) – Psycho-Pass raises important questions about criminological orientations towards technology and its role in facilitating increasingly invasive and ubiquitous forms of social control. In the spirit of Rafter and Brown’s (2011) Criminology Goes to the Movies, in this article I map out the popular criminology of Psycho-Pass. I do so not simply to identify theoretical motifs in the show, but to problematize them. Psycho-Pass, I argue, doesn’t just represent criminological theory – it also poses key thought experiments for criminologists to think about risk, security, surveillance and the nature and implications of technology.

Through examining Psycho-Pass, I bring together and analyze two phenomena that have yet to receive significant attention from criminologists: cyber-punk science fiction, and Japanese anime. Despite recent interest in comics, graphic novels and animation (see Giddens, 2015, 2017; Phillips and Strobl, 2013), there is scant literature by criminologists and legal studies scholars on such forms of popular culture emanating from non-Western regions. Like Western comic books, graphic novels (Petty, 2015) and animation, Japanese anime and manga lend themselves to analyses centring on the issues of representation, affect, aesthetics and spectatorship that visual criminology is concerned with (Brown and Carrabine, 2017). However, rather than examining the criminological aesthetics (see Young, 2009) of Psycho-Pass, in this article I instead want to unpack its dystopian vision of social control, and the questions that criminologists might take away from its depiction of a pre-crime society.

The purpose of my analysis here is, therefore, quite different to that of Wall’s (2008) analysis of social science fiction’s impact on mainstream conceptions of cyberspace. Whereas Wall was concerned with the significant impact that social science fiction writers such as William Gibson and Bruce Sterling have had on conceptualizations of contemporary ‘cyberspace’, I am interested in examining how the genre might serve as a springboard for thinking through the implications of future criminal justice technologies. In this sense, my argument is not that social science fiction should be a cultural reference point for criminological understandings of technology, but rather that it should be a jumping off point for interrogating future techno-social configurations. That is, rather than comparing representations of criminal justice in science fiction to real-life referents, we might be better served to treat social science fiction as a simulator: a machine for generating realistic imitations of potential criminal justice futures.

My argument is not that science fiction necessarily exceeds the theoretical insights offered within academic criminology. Rather, science fiction offers criminologists thought experiments for questioning and thinking through potential criminal justice futures. This is a product, I argue, of what literary theorist Suvin (1979) terms the novum of science fiction: the fictional, albeit scientifically plausible, innovation, invention and/or technologies upon which stories within the genre hinge (see Chu, 2010). Through interrogating the novum of quality social science fiction, ² criminologists are presented with a raft of questions pertaining to current and potential criminal justice technologies. Indeed, if we look back to the examples of 1984, Minority Report and Neuromancer mentioned above, it is the novum of each of these stories (Big Brother, pre-crime and cyberspace) that has inspired academic theory.

Psycho-Pass, I argue, encourages us to question a society of algorithmic tyranny – a society where the overwhelming majority of classifications are driven by algorithms. In a state of algorithmic tyranny, we are defined primarily by what Cheney-Lippold (2017: 48) terms ‘measurable types’: data templates that offer interpretations of the world, assign individuals’ identities and frequently ‘determine the discursive parameters of who we can (and cannot) be’. And in presenting a society where data speaks for us, and where crime has become a ‘measureable type’ derived solely from data, Psycho-Pass offers us a vista of what, riffing on Beer (2017) we might term the criminogenic power of algorithms. Though criminologists have recently interrogated the growth of algorithmic risk (Hannah-Moffat, 2018), the pre-crime society (Zedner, 2007a) and the challenge that Big Data presents to criminology (Chan and Bennett Moses, 2016b), these investigations are, understandably, rooted in the present. Through presenting a hypothetical society that is very much an intensification of these recently explored trends, Psycho-Pass – and other forms of social science fiction – offer criminologists a resource for thinking ‘longer term’ about the implications of these trends.

I begin this article by first discussing Japanese animation, or ‘anime’, as a site of popular criminology, along with some of the key methodological considerations of researching anime as a Western criminologist. I then provide a brief summary of Psycho-Pass, detailing both its premise and salient literary and criminological influences. Having established the key criminological motifs Psycho-Pass explores, I then examine how the show’s depiction of a pre-crime society, which takes us beyond ‘algorithmic power’ (see Beer, 2009) to full-blown ‘algorithmic tyranny’, offers criminologists a resource for problematizing the transformation of crime into a ‘data object’. I conclude by making a case for incorporating the technological imagination of science fiction into the criminological imagination. Social science fiction, I argue, offers three key resources for criminologists concerned with the crime-criminal justice-technology nexus: 1) a historico-discursive resource for archaeologically excavating and problematizing the literary inspirations of criminal justice innovations; 2) a pedagogical resource for engaging students in questions concerning the criminological implications of technological change; and 3) an analytical resource for developing both a critical consciousness of technology and the capacity to reflexively govern technological innovations in the criminal justice sphere.

Anime: An uncharted site of popular criminology

To date, studies in popular criminology have generally been Eurocentric in their subject matter, focusing on films and television shows produced in North America and Europe. This focus on the global North is not peculiar to popular criminology alone, but is symptomatic of cultural criminology more generally. In their editorial for Crime, Media, Culture’s recent Asia-style Special Issue, Fraser et al. (2017), for example, note that the administrative flavour of criminology in Asia has resulted in a dearth of cultural and critical criminological scholarship in the region (see Laidler et al., 2017). Echoing Fraser et al.’s (2017) call for greater criminological engagement with Asian media and culture, I suggest that Japanese animation – referred to in the West as ‘anime’ – represents an important, yet under-researched site of popular criminology that should be taken seriously by cultural and media criminology. Whilst anime has received significant attention in film and cultural studies (Condry, 2013), it has been conspicuously absent from criminological studies of popular media.

Before I begin my analysis, it is worth charting out some of the key methodological considerations I reckoned with in analyzing an anime franchise as a Western non-Japanese speaking criminologist. Anime producers regularly reference, remix or otherwise draw inspiration from Western popular culture, leading to genre-hybridization (Davis, 2015; Denison, 2015) and fostering the form’s global fandom. Whilst Psycho-Pass draws heavy inspiration from Western writers and theorists, ³ interpreting it through Western theory risks ignoring some of the show’s cultural specificities. Like any series, Psycho-Pass reflects the context in which it was produced. In examining how popular culture produces, mirrors and challenges criminological theory, we must, as Aas (2012) exhorts us to do, reflect on the ‘situated identity’ of these theories. In the case of Psycho-Pass, it is particularly worth reflecting on the criminological context of Japan, which gives some additional nuance to the theories referenced and/or alluded to in the show. Like its fictional counterpart depicted in Psycho-Pass, Japan is a low crime society, with one of the lowest homicide rates in the world (UNODC, 2013). Whilst Psycho-Pass’s meditations on a technologically-facilitated low crime society are certainly transferable beyond Japan (and in Psycho-Pass: The Movie, the series expands its lens beyond the country), they should be read against this backdrop of a society that already has a very low crime rate.

This article’s focus on technology, however, opened up another concern – that of ‘techno-orientalism’. Often traced back to William Gibson’s foundational work in the cyber-punk genre (Nakamura, 2013), techno-orientalism refers to the practice of imagining Asia as futuristic and hyper-technological (Roh et al., 2015; Ueno, 1999). In focusing my discussion specifically on the technological in Psycho-Pass, I do not want to repeat the techno-orientalism that Denison (2015) notes has sometimes characterized Western studies of anime. Though technology is a common theme in anime, anime extends far beyond the mecha, cyberpunk and science fiction genres it is often associated with by Western audiences. Future criminological studies of manga and anime might, for example, readily examine how incarceration, policing, organized crime and criminal justice have been represented in prison dramas such as Rainbow: Nisha Rokubō no Shichinin, and thrillers such as Monster and Death Note.

‘I’ve only ever obeyed that gun’s orders’: A brief summary of Psycho-Pass and its criminological parallels

We live in a time when man’s own mind is wide open to anyone with the right machine. (Tomomi Masaoka, Psycho-Pass)

Before beginning my analysis, a brief summary of Psycho-Pass is required for readers unfamiliar with the series. Described by Pinchuk (2016: N.P) as ‘Minority Report meets a Clockwork Orange by way of Ghost in the Shell’, Psycho-Pass depicts a 22nd century Japan under the Sibyl System: a network of psychometric scanners that assess not only the likelihood of an individual committing a crime, but also their aptitude for various roles and occupations in society. Through following protagonist Akane Tsunemori, a member of the Criminal Investigation Department (CID) of Tokyo’s Public Safety Bureau, the show examines the moral, social and cultural implications of a society governed by pre-crime, ubiquitous surveillance and algorithmically-imposed identities. As explained in the first volume of the series’ prequel manga, Psycho-Pass is set in:

[a] future, where one’s mental state and personality can be measured into numbers. Every emotion, desire, social deviation, and mental inclination is put into record … This value that determines the standards of human mind, even an individual’s soul itself … is known to the public as ‘psycho-pass’. (Sai and Gotou, 2017: 6–7)

In the world of Psycho-Pass, public and private space have both been transformed, to use Kitchin and Dodge’s (2011) term, into sophisticated ‘code spaces’ where individuals are intermittently surveilled by a network of psychometric scanners monitoring their psychological wellbeing and the probability that they will commit a crime. This last metric is referred to in the show as an individual’s ‘crime coefficient’: the component of their ‘algorithmic identity’ (Cheney-Lippold, 2011, 2017) that quantifies their risk of engaging in harmful or illicit behaviour. When an individual’s ‘crime coefficient’ passes a threshold, it alerts members of the CID, who are tasked with apprehending them before they commit a crime, or, if their psycho-pass crime coefficient is high enough, summarily executing them.

The Sibyl System’s role in ‘preventing’ crime doesn’t end with surveillance, however. Members of the CID are armed with algorithmically-driven guns called ‘Dominators’ that only fire when the target’s crime co-efficient exceeds an acceptable level. When aimed at a target, a Dominator reads and sends continuous psycho-metric data about the individual to the gun’s user. If the targeted individual’s crime co-efficient exceeds a certain threshold, the gun’s user is then instructed by the weapon to fire at the individual. The Dominators, then, are much more than merely ‘tools’ employed by agentic law enforcement officials. Instead, they might more accurately be described as technologies that employ and direct CID agents, imposing a logic of actuarialism on their use of force. Indeed, in one scene, CID enforcer Tomomi Masaoka even states that ‘he follows that guns’ orders’. As the eyes of Sibyl and a direct expression of the state’s will, the Dominators offload choice, decision making and judgement from the CID agents wielding them onto technology. In this way, the Sibyl System – and the artificial intelligence that seemingly drives it – functions as judge, jury, executioner and watcher.

As the above description shows, Psycho-Pass has numerous criminological parallels. Perhaps most saliently, the show’s overt references to surveillance, actuarial risk assessment (see Feeley and Simon, 1992, 1994) and the scientification of social control resonate with many classic ‘surveillance essays’ (see Marx, 2016). These include Deleuze’s (1992) essay on the rise of societies of control, Cohen’s (1985) work on ‘net widening’, Clarke’s (1988) work on dataveillance and, of course, Foucault’s (1977) work on the disciplinary society and panoptic regimes of surveillance (see Santy and Soelistyo, 2014). Indeed, the panopticon is explicitly discussed in the second season of the show as the name of a potential replacement to the Sibyl System. Though the Sibyl System is certainly panoptic in its normalizing aims, it is more akin to what Poster (1990) terms the superpanopticon, a ubiquitous panopticon ‘without walls’, than it is to Foucault’s (1977) original conceptualization of bounded panoptic surveillance in prisons, schools and factories. Within the fictional Japan of Psycho-Pass, citizens are almost constantly enveloped within Sibyl’s ‘surveillant assemblage’ (Haggerty and Ericson, 2000), which, drawing on biometric and psycho-metric observations (see Aas, 2006), generates a digital ‘data double’ of their calculated stress levels and probability of committing a crime. Yet despite having the clearest parallels to recent criminological work concerning the impact of digital technologies on criminal justice, many of the issues raised by Psycho-Pass have a long history within criminological thought. Sibyl’s aptitude tests, for example, recall early 20th century eugenic criminologist Henry Goddard’s (1920) recommendation that all members of the population be administered mental tests and assigned jobs on the basis of their results (see Rafter et al., 2016). Further, the notion of ‘latent criminals’ – individuals who (due to their psycho-biological makeup) are deemed by the Sibyl System to have a high probability of committing a crime – recalls Lombroso’s (2006 [1876]) notion of the ‘born criminal’.

Whilst Psycho-Pass references numerous works of science fiction, ⁴ the show is particularly indebted to the work of science fiction author Phillip K Dick, whose dystopian stories have been adapted into a number of blockbuster Hollywood films including Blade Runner, Total Recall, Minority Report, A Scanner Darkly and The Adjustment Bureau. ⁵ Dick’s influence on Psycho-Pass is most pronounced in the Sibyl System, which illustrates perfectly his notion of pre-crime: strategies used by law enforcement agencies that attempt to stop crimes before they are committed. The ability to forecast crime, is not, however, that far-fetched from current policing practices. Originally introduced in Dick’s (1956) short story ‘The minority report’, the notion of pre-crime has increasingly been taken up by criminologists to explain the rise of strategies within law enforcement and security agencies that bring the notion to life (Zedner, 2007a, 2007b). Yet whilst the Precrime department of Dick’s Minority Report is the product of magic – namely, a trio of ‘mutants’ with the power of precognition – the Sibyl System is, ostensibly, a technological marvel; a product of science, rather than the supernatural.

In this way, Sibyl can be read as a more advanced and invasive form of many contemporary pre-crime initiatives, from predictive/algorithmic policing (see Chan and Bennett Moses, 2016a; Van Brakel and De Hert, 2011) to CCTV surveillance that uses algorithms to either identify offenders (Introna and Wood, 2004) or predict when violence is about to occur in an area (O’Callaghan, 2015). Interestingly, though, whereas much of this literature on actuarial risk and algorithmic risk focuses on the ‘background’ factors that determine an individual’s risk profile (socio-economic background, past offending history etc.), Psycho-Pass presents us with a pre-crime intervention that is, to put it in Katz’s (1988) terms, just as focused on the emotional foreground of crime: the thought processes and emotions that precede and occur during a crime.

Taken together, the fictionalized Japan of Psycho-Pass amounts to what Zedner (2007a) terms a pre-crime society, ‘a society in which the possibility of forestalling risks competes with and even takes precedence over responding to wrongs done’. ‘In a pre-crime society’, Zedner (2007a: 262) explains, ‘there is calculation, risk and uncertainty, surveillance, precaution, prudentialism, moral hazard, prevention and, arching over all these, there is the pursuit of security’. ⁶ Under Sibyl, which is named in reference to another of Dick’s (1987) works, The Eye of the Sibyl, ⁷ Japan is a country that is, to use Valverde’s (2001: 83) phrase, governed through security: each citizen’s risk of committing a crime is calculated through ubiquitous state surveillance (see Schinkel, 2011). Following a preventative and prudential logic, citizens whose risk level exceeds a certain threshold and who are deemed a moral hazard are sent to ‘emergency therapy’, or permanently removed from circulation through summary execution.

The criminogenic power of algorithms

Up to this point, I have detailed some of the key parallels between Psycho-Pass and criminological theory. In this section, I hope to address another issue: where does Psycho-Pass help take us beyond existing criminological knowledge? For whilst it is one thing to note the pedagogical value of science fiction for criminologists in representing existing issues pertaining to technology and criminal justice, it is quite another to argue that the genre offers criminologists thought experiments for thinking longer term about technology. In what follows, I hope to demonstrate that Psycho-Pass does not just mimetically represent already knowable issues pertaining to technology and justice. Rather, it poses new questions and directions for criminologists concerned with the criminal justice-technology interface.

As a form of popular criminology, the twin novums of Psycho-Pass – its pre-crime Sibyl System and the Dominator weapon’s wielded by its main characters – bring to bear important questions for criminologists concerned with the contemporary technological unconscious, ⁸ and the potentialities stemming from its logic of interoperability and networked interconnection. Psycho-Pass, in short, presents a future world governed by a technological unconscious with an overactive superego. And though the show probes a number of issues relating to the increasingly salient superegoic dimension of the technological unconscious, two stand out: its illustration of the criminogenic power of algorithms and its problematization of crime as a ‘measureable type’ and data object.

Throughout much of the first season, the nature and workings of the Sibyl System are not revealed to viewers, nor known to the characters themselves. Like most digital technologies, its workings are opaque; it is trusted, but not understood by those who interface with it. Despite the lack of understanding regarding how it functions, the Sibyl System is trusted by CID officials and most members of the population, so much so that on several occasions characters feel compelled to follow its directives, even in instances where they entail clearly unjust outcomes. In the opening episode of the series, for example, rookie CID inspector Akane Tsunemori is called to her first case: a man who, knowing he will soon be arrested, has abducted a woman after receiving a high crime co-efficient rating. After sexually assaulting and torturing the woman, the man is summarily executed by one of the CID agents upon orders from the Sibyl System. The psychological distress caused by the event, however, causes the woman’s own crime co-efficient to spike, leading the Sibyl System to direct the inspectors to fire a non-lethal shot at the woman and take her in for ‘emergency therapy’. After a further spike in her crime co-efficient, the Sibyl System directs the CID members to ‘eliminate’ (kill) the target and switches their Dominators to ‘lethal elimination’ mode. Tsunemori ultimately disobeys Sibyl’s directives, opting to instead pacify her subordinates and wait until the woman’s ‘threat level’ has been reappraised before firing a non-lethal ‘paralyzer’ shot to subdue her. Only through doing so is Tsunemori able to, in the words of fellow protagonist Shinya Kogami, ‘put justice before duty’.

This episode, along with several others over the course of Psycho-Pass’s two series, vividly illustrates the criminogenic power of algorithms. Through labelling individuals as ‘at risk’ of offending or placing them within other ‘measureable type’ categories that tangibly affect their lives, machine-learning AI, like actuarial risk programmes before them, have the power to cause criminal behaviour. This criminogenic power of algorithms might be superficially understood through the terms of classic labelling theory: a machine-learning algorithm identifies an individual as matching their dynamic data template of ‘crime’; the individual internalizes this algorithmically inferred assessment; as one of many potential responses to this labelling process (Rogers and Buffalo, 1974), they engage in secondary deviance catalyzed by the very act of being labelled (Lemert, 1967).

Indeed, on several occasions throughout the series it is the act of being labelled a threat by Sibyl that precipitates a character’s descent into criminal activity. This alone is not particularly novel within criminology; many secondary crime prevention programmes have been criticized for their counterproductive effect of labelling the populations they target as ‘at risk’ of offending (see Sutton et al., 2013). What difference does it make, then, that it is a machine, rather than a human, that is doing the labelling? Firstly, when aligned with a pre-crime rationality, machine-learning algorithms may lead to the paradoxical situation of secondary deviancy without primary deviance. In Psycho-Pass, false-positive classifications by Sibyl abound. Individuals who pose no actual risk of committing a crime are on multiple occasions scheduled for ‘emergency therapy’ by Sibyl – an act that leads some to a complete overhaul of their self-identity. As the next section will explain, this is a product of crime’s datafication – its transformation into a data object and ‘measureable type’ – and the datafictions that are produced by and sustain this process.

Crime as ‘data object’: Or, datafication and its datafiction

‘A measureable type’, Cheney-Lippold (2017: 47) explains, ‘is a data template, a nexus of different datafied elements that construct a new, transcoded interpretation of the world’. As ‘datafied versions of ideal types’ (Cheney-Lippold, 2017: 51), they also accentuate certain dimensions of a phenomenon (Weber, 1978). Further, to model measureable types, AI must quantify dynamic phenomena. What constitutes crime and risk as ‘measurable types’ is not static, nor solely at the whim of human programmers. Rather, through machine learning, the technological unconscious is able to assimilate new markers of crime and risk from the data it constantly collects. Resonating with recent work in the philosophy of information (Floridi, 2013), Psycho-Pass explores the dangers of crime becoming a measurable type and a ‘data object’. As Cheney-Lippold notes, we must distinguish between ‘terrorist’ (or criminal) as an object of legal and political knowledge, and terrorist (or criminal) as a data object. For though an individual who commits an act deemed terroristic may retroactively be classified (legally and politically) as a terrorist, they may not, according to the machine-learning algorithms that seek to identify such individuals, constitute a ‘terrorist’ as a data object.

This profound limitation of crime as a measureable type is dramatized in Psycho-Pass by its season one antagonist, Shogo Makishima – a highly intelligent and alienated individual who seeks to destroy the Sibyl System and the society it has created. Makishima is unable to be apprehended through the very system he seeks to destroy; classified as an individual who is ‘criminally asymptomatic’, he is able to keep his crime co-efficient low, meaning that CID are unable to apprehend or kill him using their Dominator weapons. Psycho-Pass, in other words, presents a world where crime as a socio-cultural and political category has been almost completely superseded by ‘crime’ as a data object, so much so that the state is impotent in the face of violent individuals who the Sibyl system does not classify as criminal.

Demonstrating the limits of Sibyl’s ‘technological solutionism’ (see Morozov, 2013), Psycho-Pass therefore depicts a world where AI-informed pre-crime is always afflicted by false-negatives. Yet, despite this shortcoming, the Sibyl System is, like many real-life criminal justice technologies, lauded by its proponents as neutral and objective – a means of counteracting the foibles of human decision making. Psycho-Pass’s Sibyl System therefore illustrates how algorithms often provide a facade of objectivity to value-laden criminal justice practices. To use Ferrell’s (1993) term, they offer a compelling ‘aesthetic of authority’ for criminal justice practitioners to legitimate their actions. This veneer of neutrality is repeatedly questioned throughout the series and finally shattered when the nature of the Sibyl System is revealed. Towards the end of the first season, villain Makishima discovers that, far from being driven by software, the Sibyl System is a wetware hive mind of (‘criminally asymptomatic’) people whose brains have been surgically exercised and subsequently housed in a special container. Upon being informed of the true nature of Sibyl, Makishima notes that trust in the system was predicated upon its ostensible impartiality:

What a farce. They promised a society managed by fair and impartial machine intelligence. A law free of petty human ego. People only accepted Sibyl because that’s what they thought they were getting. But it was a lie – Sibyl’s arbitrary and unfair. It’s nothing but a committee of pickled brains in mason jars. (Makishima, Psycho-Pass)

The revelation that the Sibyl System is a hive mind rather than AI hides the truth that algorithms are, to use Latour’s (1990) phrase, already ‘society made durable’. Further, it reveals a central truth about algorithmic regulation and Big Data: that datafication requires datafictions to gain and obtain legitimacy. Like Big Data, algorithmic regulation is a cultural as well as a technological phenomenon (Boyd and Crawford, 2012), and as such it requires its own ‘mythology’ and faith in order to function.

Psycho-Pass’s central conceit is that criminal justice technologies are never impartial – they reflect the values and presuppositions their creators hold and programme into them. What the Sibyl System illustrates is that the technological unconscious is, to use Jameson’s (1981) term, a political unconscious of values inscribed into technology. Despite its ‘aesthetic of authority’, raw data collected and analyzed by digital surveillance technologies has no inherent meaning, but is given meaning by programmes developed by humans. As Marx (2016: 49) notes, ‘the signs, markers, emanations, tracks, traces, remnants, deposits, debris, and residuals that accompany being in the world and interacting with others are given new meanings by technology’.

More than just exploding the myth that technologies may be completely objective or neutral (see Feenberg, 2002), Psycho-Pass offers a cautionary tale on what might be lost if crime is transformed into a data object. In short, the Japan of Psycho-Pass is a society of algorithmic tyranny: a society where, to quote Cheny-Lippold (2017: 25), ‘data about who we are becomes more important than who we really are or who we may choose to be’. In a society of algorithmic tyranny, who gets classified as a criminal is entirely determined by algorithms: it is a citizen’s datafied self or ‘data double’, rather than their biographical self, motivations and desires, that is judged. A society of algorithmic tyranny is, in other words, post-narratival: data speaks for us, and patterns stand in for explanation (see Chan and Bennett Moses, 2016b). Our data, and the algorithmically-inducted identities that are inferred from it, drive our lives and opportunities, whilst narrative, context and everything else that cannot be datafied and assimilated into ‘measureable types’ are removed as indicators of risk. In the world of Psycho-Pass, such algorithmic tyranny has a hefty price. Characters are seen constantly checking their own Psycho-Pass readings, afraid that they may be labelled a ‘latent criminal’ by Sibyl and targeted for emergency therapy. Biographical self is driven by datafied self. Further, in placing faith in technology and allowing crime to become a data object, criminal justice is, at least ostensibly, turned over to an acritical form of algorithmic regulation that not only erases context, politics and narrative, but also supplants human knowledge and decision making.

Conclusion: Science fiction and the criminological imagination

Psycho-Pass is, as Sakurai (2015a, 2015b) puts it, an ‘ambiguous dystopia’. Most of the citizens within Psycho-Pass’s fictionalized Japan are, arguably, happier under the Sibyl System than they were before it. Under this façade, though, is a grimmer reality: that by all but eliminating one form of violence through pre-crime, you only increase another – the structural and symbolic violence of the state, which comes to have complete totalitarian control over every single citizen. So utopian is the concept and practice of pre-crime that it can quickly turn into a very dystopic vision of social control. At its core, then, Psycho-Pass depicts a profoundly utilitarian vision of social control, where the (ostensible) happiness of the many is secured by the misery of a few.

As Raymen (2017: 1) notes, dystopian fiction offers criminologists ‘an opportunity for a popular criminology to address what criminologists have described as our discipline’s aetiological crisis in theorizing harmful and violent subjectivities’ (see also Yar, 2015). In addition to agreeing with Raymen, I argue that dystopian science fiction also offers an opportunity for popular criminology to address harmful, unjust and/or criminogenic criminal justice technologies. Psycho-Pass and social science fiction offer very different dystopian visions to the post-apocalyptic fiction discussed by Raymen. Whereas post-apocalyptic dystopia can, as Raymen (2017) demonstrates, be read viz-a-viz Žižek (2010) as a fetishistic disavowal of the violent subjectivities produced by late capitalism, Psycho-Pass and other cyber-punk fiction can more readily be read as cautionary tales in technology. In this sense, the cyber-punk of Psycho-Pass is diametrically opposed to that of the post-apocalyptic dystopia of The Walking Dead. For whilst the latter arguably acts to render the violent core of capitalist subjectivity unconscious, the former might, like much other science fiction, instead be said to promote a critical consciousness of technology’s unconscious (Wood, 2017).

Drawing again on Miller and Bennett (2008), I would like to conclude by quickly delineating three ways in which science fiction may play a role in shaping the ‘criminological imagination’ (Young, 2011). Science fiction, I argue, offers criminologists at least three key resources: 1) archaeological resources for tracing the literary inspirations behind criminal justice technologies; 2) pedagogical resources for promoting a critical consciousness of technologies among students of criminology; and 3) capacity building resources for criminologists to reflexively understand and govern new criminal justice technologies.