The technological veil conceals the reproduction of inequality and enslavement.
—Herbert Marcuse, One-Dimensional Man
It is tempting to view the explosion of criminal justice technology in the 1990s as inevitable given the digital revolution. But the inordinate amount of resources that have gone into designing, developing, and deploying these technologies cannot be explained by the engineering laws of Gordon Moore or Mark Kryder. The flight of industrial capital, racialized deproletarianization, social struggle, the crime and drug wars, and the rise of information capital each played a role in triggering the criminal justice system’s multi-billion-dollar digital metamorphosis. Since the final quarter of the twentieth century, northern industrial cities have undergone a slow and uneven shift from manufacturing sites into sites of conspicuous consumption, financial exchange, real estate development, and IT industry. Some prime features of this shift included the privatization of public services, criminalization of poverty, and the ascent of information infrastructure. The digitization of the carceral state cannot be understood apart from these developments.
For city governances, large-scale projects to establish digital infrastructures for criminal justice were rooted in efforts to manage contradictions specific to postindustrialized urban space. On one hand, the forces of global economic power bunkered inside urban arcologies made up of banks, condominiums, financial firms, gentrification, high-end shopping, and tax abatements. Urban administrations, no matter their party composition, funneled development capital, tax concessions, and zoning bonuses to various industries characterizing upstart global cities. Urban decision-making was slowly but steadily handed over to business improvement districts, special purpose development corporations, and tax increment financing districts insulated from public oversight. Urban planning initiatives were quietly steered by technology corporations, such as CISCO, IBM, and Siemens, that charter IT-friendly “solutions” for architects, city planners, engineers, and public officials in charge of cities.
On the other hand, the “territories surrounding these nodes play an increasingly subordinate function.” Indeed, the formation of securitized centers of urban wealth went hand in hand with the formation of deindustrialized, devalued neighborhoods. Upscale lofts came to coexist with dilapidated housing projects, and the latter spread at extraordinary rates. Between the 1980s and 1990s, sociologist William Julius Wilson calculated industry jobs for workers without high school degrees decreased by 73,000 in Baltimore, 89,000 in St. Louis, 172,000 in Philadelphia, and 492,000 in New York City. Public workforces were also significantly downsized. Between the 1970s and the 1980s, unemployment in some predominantly black neighborhoods increased at quadruple the rate of city averages, and in some cities, as many as 80 percent of black families lived in poverty. As the outmoded labor force’s economic function declined, so too did the state’s inclination to support its social reproduction. Cities hemorrhaged affordable housing, as contractors prioritized high-income housing to abet gentrification over low-income housing for obsolete workers. In 1970, there was a surfeit of almost 2.5 million low-income housing units in the United States. By 1985, there was a deficit of 3.7 million. From 1978 to 1988, the Department of Housing and Urban Development’s appropriations for subsidized housing were slashed by 70 percent. Homeless populations expanded, which were disproportionately made up of female victims of domestic violence, displaced elders, the mentally ill, and undocumented immigrants. Mass criminalization was one solution to managing the political radicalism and social problems that sprang from these conditions.
All of these mutations took place alongside the revolution of information capital. On the technical side of things, this world-changing event was made possible by the combination of the internet, microprocessors, personal computers, relational databases, and the World Wide Web. For the business sector, the digital landscape established new horizons to conduct exchange, segment consumer markets, and optimize logistical chains. New profit margins were discovered through data mining using Gregory Piatetsky-Shapiro’s knowledge discovery in databases techniques. Businesses wielded these techniques for everything from employee scheduling to micromarketing to price strategizing, among other practices. Knowledge industries that emerged from these developments were centered in cities: Cambridge, Raleigh-Durham, San Francisco, San Jose, Seattle, Washington, D.C. In fact, the steady expansion of urban tech hubs exacerbated the obsolescence of the very populations whom IT companies eventually designed technologies to digitally carceralize.
Seeds of Police Computerization
Geographic information systems were crucial to police initiatives to incorporate computer technology during the 1960s. The systems, which manage spatial data, were trumpeted by police officials and software companies as means of patrolling criminalized areas with greater cost-effectiveness and efficiency. Spatial data were touted as being especially valuable for police units dispatched to suppress urban uprisings. Political radicalism was on the offensive. As the 1960s raged on, civil rights and black nationalism drew closer to Third World and labor movements, spreading their influence through Chicanx, feminist, indigenous, and queer struggles. Speaking in 1964 in front of the Militant Labor Forum, Malcolm X argued that “America’s strategy is the same strategy as that which was used in the past by colonial powers: divide and conquer.” In the political field, this strategy involved gerrymandering, sabotage, and the installment of a police state. “Algeria was a police state,” X observed. “Any occupied territory is a police state; and this is what Harlem is. . . . The police in Harlem, their presence is like an occupying army. . . . They’re in Harlem to protect the interests of the businessmen who don’t even live there.” This radical interpretation of racialized governance as a mode of internal colonialism was foundational for the Black Panther Party for Self-Defense, which was forged to establish a human infrastructure that included, among various services, the armed protection of civilians from the Oakland Police Department following a bloody rebellion in Watts, California, in 1965. One way that the state responded to the surge of black radicalism involved restructuring its enforcement apparatuses. The most totalitarian expression of these reactions was the Counterintelligence Program (COINTELPRO), infamously devised to “expose, disrupt, misdirect, discredit, or otherwise neutralize the activities of black nationalists.” COINTELPRO was handled by cold warriors whose conception of policing technology was mostly limited to wiretaps. But near the same time, more sophisticated technologies began to appear at the state level.
The introduction of geographic information systems into the police apparatus was to a great degree a function of happenstance. Geographic information systems first appeared in police departments in the 1970s, alongside the quantitative revolution in the discipline of geography. This significant disciplinary development catapulted spatial science to the forefront of academic spatial knowledge production. The revolution was led by a post–World War II generation of geographers seeking to establish a universal ontology of social space through descriptive statistics, inferential statistics, probability theory, and social physics. This new perspective, in which human geographies were seen as calculable, predictable, and controllable, was deeply appealing to police departments leading the crime and drug wars. The first attempt to fuse quantitative geography with policing was undertaken in the early 1970s by Phillip S. Mitchell, an expert in mathematical programming and public systems working for the California Council of Criminal Justice (CCCJ). An associate of the International Association of Chiefs of Police and board member of the CCCJ’s Research Development Task Force, Mitchell pushed for using spatial clustering methods to decide where police should concentrate patrol forces. Mitchell’s work was just a few years removed from the founding of geographic information software companies Esri and M&S Computing (now Intergraph). Mitchell was perhaps too far ahead of his time. The cost of mainframe computers and relative lack of updated geographic base files during the 1970s limited digital police mapping from spreading to most departments for a few decades. Nevertheless, Mitchell’s work was foundational in bringing differential policing into programming language. With Mitchell, the rationalization for concentrating patrol forces in already criminalized areas was expressed as
where p represents the expected number of incidents of type m based on past events, t(kʹ) the average driving time required to respond to an incident for patrol unit k, and S(kʹ) the average incident load for the unit. Equations like these signified the dawn of a new conception of urban space. The city was rendered a grid of rationally distributed patrol forces (Figure 9). Of course, the idea that the police apparatus should be the choice instrument to manage social problems and struggles was itself a political invention. But where southern segregationists normalized racialized policing through biological racism, explicit laws, and overt racial demagoguery, northern powers normalized it through mathematical formulas.
The geographic information industry slowly expanded during the 1980s, which saw the establishment of companies such as Autodesk, GE Energy/Smallworld, and MapInfo. Different state apparatuses, such as the U.S. Defense Mapping Agency, played a role in assisting these companies. Founded in 1972, the mapping agency was a Vietnam War invention for generating digital maps, charts, and positional data for military operations and weapons systems. The emphasis on spatial data production reached the domestic police apparatus in the face of mounting opposition to police violence brought about by Reagan’s War on Drugs. Antagonism between police and the negatively racialized poor became increasingly volatile and reminiscent of Fanon’s description of a colonial “compartmentalized world.” Black separatism began to reappear through a revitalized Nation of Islam. Cedric Robinson published Black Marxism in 1983. Hip-hop music morphed into ethnographic analyses of racial criminalization, punctuated by Niggaz Wit Attitudes’ (NWA) 1988 billboard success “Fuck the Police” and Ice T’s “Cop Killer” four years later. These tensions boiled over following the acquittal of Los Angeles Police Department (LAPD) officers who brutalized Rodney King while being recorded in 1991, which detonated revolts on a massive scale. Fanon could have easily predicted what happened next: the result was not a rational confrontation of viewpoints but rather some seven thousand federal troops deployed to suppress popular insurrection, twelve hundred of whom were Marines. Approximately one thousand people lost homes, twenty-four hundred were injured, twenty thousand lost jobs, ten thousand were arrested, and $1 billion worth of capital was destroyed.
As computing became cheaper and more powerful, urban police departments turned to geographic information systems in efforts to convert insurgent spaces into rationally managed microspaces. The practical effect was to mark off stigmatized territories, classify them according to varying levels of policeability, and codify uneven distributions of legalized violence with the aid of spatial statistics. Geographic information systems provided technical support for these objectives. The federal government subsequently poured funds into the Institute for Intergovernmental Research, International Association of Directors of Law Enforcement Standards and Training, and National Training and Technical Assistance Center to develop its spatial data infrastructure. Drawing inspiration from the blossoming subdiscipline of automated geography in the academy, these collaborations unleashed a series of projects to integrate geospatial visual analytics, spatial econometrics, and three-dimensional simulations of cities for police.
Programs to reorganize the police apparatus around geographic databases spread like wildfire as the millennium’s end approached. In the 1990s, the National Institute of Justice (NIJ) launched its Crime Mapping Research Center (CMRC) in the behavioral sciences division. CMRC teamed with Esri to create a suite of software including a free application that enabled police to record geocoded information and display it on digital maps, crime analysis extensions for standard geographic software, and regional crime analysis application enabling information sharing across police jurisdictions. Five years following its launch, the CMRC was renamed Mapping and Analysis for Public Safety (MAPS), and it shifted its focus to developing artificial intelligences to find new types of patterns in police geographic systems. This digitized production of knowledge was not intended to deepen understandings of criminal behavior. Police mapping was rather meant to produce logistical knowledge to suppress urban radicalization and social disorders. Such knowledge arose amid city efforts to manage the racial and spatial contradictions inherent to the accumulation of capital. As Henri Lefebvre discovered, capital overcomes crises by producing new types of spaces—not only material space but also representations of space. In the case of policing, legions of city authorities, computer scientists, criminologists, and software firms created a boundless variety of cartographies to aid patrol units: raster maps depicted densities of reported incidents through shaded census tracts, police precincts, and zip codes; density maps subdivided urban space according to reported incident rates; kernel density maps displayed reported incident densities in a continuous fashion that superseded administrative boundaries; standard deviation ellipses differentiated hot spots according to their size; and kernel smoothing represented hot spots through a type of geocoded thermal vision. What these representations of space held in common was that they all mapped something inhibiting urban redevelopment—the growing mass of devalued populations. And whereas the maps of social ecology begat archetypal zones, those of computerized criminology begat targets.
Heat mapping was only one aspect of police computerization; the latter also involved populating databases with information about criminalized populations. In 1987, the Los Angeles Sheriff’s Department and Law Enforcement Communications Network introduced the Gang Reporting, Evaluation, and Tracking system, a database of two hundred thousand purported gang members and affiliates. The database, later renamed CalGang, included more than two hundred data elements about suspects (addresses, acquaintances, gang monikers, names, physical descriptions, photographs, Social Security numbers) and gangs (racial makeup). By the mid-1990s, 40 percent of incarcerated juveniles, 25 percent of incoming male inmates, and nearly 15 percent of nonaligned inmates were registered in gang databases. At the time, somewhere around 70 percent of large police departments were using gang databases as well. Though legal definitions of gang member often are not written into state legislative codes, they are nevertheless written into computer code. Individuals observed in photographs with known gang members or who have corresponded with known gang members, have gang tattoos, or write about gangs in books, in papers, or on walls are susceptible to being categorized as gang members.
Between 1995 and 1996, Congress appropriated $67.5 million of defense funding to transfer military technology to police departments. The Virginia-based Rapid Prototyping Facility, managed by the FBI and Advanced Research Projects Agency, designed an automobile tracking system made up of a grid of hidden sensors strewn throughout a city that tracks vehicles in real time. The 1990s also saw the government and technology companies explore ways to cybernetically enhance patrol units. In an effort to compress urban space-time for patrollers, police turned to the fledgling field of mobile computing during the War on Drugs. Automated vehicle locators coupled with portable terminals were first used to match emergency calls with patrol units in the 1970s. But moving into the following decade, companies including Motorola and Tyler Technologies began entering into the computer-aided dispatch systems market. These firms demonstrated how patrol officers could access local and national criminal history databases and databases in other bureaucracies. By the late 1990s, the Texas Transportation Institute had prototyped the first smart patrol car, equipped with touchscreens providing access to GPS vehicle locators, radar, and video feeds. This was the beginning of the transformation of the patrol unit into a miniature and mobile control room. The idea was for officers to sit behind neon screens as they navigated the criminalized territories. This meshing of officer and machine found its subjective expression in technologies devised to enhance the “situational awareness” of patrollers. Situational awareness was a martial theory about relations between human perception and environments. It is anchored in the notion that the lack of environmental knowledge is the root cause of human errors. Integrating digital computers into these activities, according to technocrats and IT companies, could simultaneously increase human environmental knowledge while decreasing mental workloads. The emergence of situational awareness in the policing domain was a boon for information capitalists. Technology companies like L-Tron Corporation designed webinars to help police departments minimize potential distractions for beat patrols; corporate–academic associations such as the International Symposium on Wearable Computing and consulting companies such as the New York State Technology Enterprise Corporation designed augmented reality training programs where officers interact with three-dimensional simulations and wearable technology. Large telecommunications companies like Verizon and small ones like WirelessUSA Inc. rebranded old information communication technology as situational awareness enhancers. The Violent Crime Control and Law Enforcement Act (1994) provided broad support for cutting-edge identification systems, criminal records systems, fingerprinting systems, and mobile digital terminals to further abet digitally enhanced situational awareness (see section 210501 of the act). In 1998, the Clinton administration signed the Crime Identification and Technology Act, which authorized $1.25 billion in state grants to develop information and identification technologies so that patrol units could gain better awareness of who occupied the environments they patrolled.
The large-scale production of policing technologies brought the state, technocrats, and the IT sector closer with respect to mass criminalization. By the end of the 2000s, teams of criminologists, computer scientists, police personnel, and state officials had established journals, quarterlies, and websites to address police technology issues ranging from tutorials for software to tips for implementing systems. Crime science for the digital age was born. The decade also saw the federal government fund private policy consultants to develop crime mapping analysis software to be provided freely to police departments across the country. Police began to produce and circulate journals espousing the virtues of affordable commercial GIS packages, SQL compatible databases, and statistical software. The Police Foundation introduced a quarterly forum for crime cartographers to keep abreast with “timely topics that are well written and practical for application—perhaps even written by some of you reading this column!” Magazines such as Government Technology, Law Enforcement Technology, and Law Enforcement Product News appeared to exploit the technofetishism that permeated police circles. Academic journals such the Journal of Quantitative Criminology and Crime Science emerged under the banners of “crime forecasting” or “prospective hot spotting” for the expressed purpose of assisting police in adapting computer technology. This lucrative enterprise became a central site in reproducing the symbolic violences inherent to racial capitalism as it upheld a legacy of restricting the legibility of the racially stigmatized urban poor to signifiers of deviance and punishment. Thus the computerization of the police apparatus was key in broader projects of urban administration to turn deproletarianized minority neighborhoods into administrative units characterized by, to borrow from Charles Pinderhughes, a “geographically-based pattern of subordination of a differentiated population.” Although the era of digitized policing was rolled out under the pretext of precision targeting, it spread alongside the increasingly differential and invasive policing of these populations in their entirety.
The introduction of digital computing to the correctional apparatus simultaneously centralized and decentralized the production of carceral space. This contradictory process was spurred by the responses of city authorities, federal technocrats, and a nascent IT sector to prison overpopulation, the prisoner’s rights movement, and the Great Recession. On the quantitative side, any number of facts might be recited to cast into relief the sheer scale of the carceral project to contain these urban problems. From the 1980s to the mid-1990s, the prison population increased threefold, correctional spending increased fourfold, and drug convictions increased almost tenfold. So racialized was the carceral explosion that black males ended up 6 times more likely to be incarcerated than white males, latinx males were 2.3 times more likely to be incarcerated than white males, and indigenous youth were 3 times more likely than white youth to be held in juvenile detention. The total number of black women incarcerated eclipsed that of any male population in other Western capitalist nations. And although the United States accounts for only 5 percent of the human population, it came to account for nearly a quarter of its prisoners.
On the qualitative side, old forms of stigma were reactivated and adapted to the times to normalize the rise of the carceral state. Social scientists were key in this regard. Urban black youths were cast by criminologists, political scientists, and public officials as “superpredators” belonging to an underclass that posed an existential threat to the whole of civilization. Such dehumanization was perhaps necessary given the inhumanity of the carceral apparatus. The barbarity of this apparatus was made abundantly clear in a 1999 report by Amnesty International, Not Part of My Sentence: Violations and Human Rights of Women in Prison. The report found that more than eighty thousand incarcerated women were mothers who were separated from more than two hundred thousand minors. By the late 1990s, some twenty-two hundred pregnant women were in jail, and more than thirteen hundred babies were born to incarcerated mothers. What is more, each state established laws that allowed for the termination of carceral captives’ parental rights. And in many instances, this cruel machine exacerbated the everyday violences experienced by criminalized minority women. Amnesty International found that black and latinx inmates in New York—the majority of whom were incarcerated for drug-related reasons, minor property crimes, or public order offenses—had significantly higher likelihoods of experiencing domestic violence, incest, and rape than the wider population.
The introduction of computer technology into correctional facilities was jump-started by a surge of organized resistance to these conditions. The first systematic programs to computerize corrections took place during the 1970s, which witnessed the prison population increase by nearly 50 percent. The political dimensions that underwrote prison growth during this decade were captured in Black Panther Party member George Jackson’s prison letters, Soledad Brother, which chronicled the politicization of incarceration following the urban revolts a few years prior. Jackson documented in vivid detail the crippling effect that politicized prisons had on the Black Power movement, American Indian Movement, and labor struggles made by rural white workforces. Shortly after Soledad Brother appeared, the Manhattan House of Detention (MHD) was jolted by resistance. Though MHD was designed to hold 925 inmates, it held more than 2,000. It was so overcrowded that almost half of its one-person cells contained three inmates. Inspections found that the complex was grossly underfunded, which resulted in shortages of beds, medical accoutrements, and soap, and frequently had scourges of rats, roaches, and lice. In August 1970, inmates detained five prison guards, demanding the city respond to inedible food, brutality, overcrowding, racism, and indefinite pretrial detention periods. Several participants in the MHD revolt were transferred to Attica Prison, which erupted in 1971 into the revolt of one thousand prisoners early in September. The Attica Prison Liberation Faction emerged at the forefront of the rebellion to issue the state with a manifesto of demands that included, but were not limited to, adequate medical attention, constitutional protections, minimum wages for prison labor, democratically elected parole boards, the right to legal representation at parole hearings, more sanitary conditions, and ending punishments on the basis of political views. The manifesto captured the radical subjectivities generated by revolt in Soledad Prison, Folsom Prison, and San Quentin around the same time. Administrative brutality was the most visible aspect, though not the only aspect, of the state’s response to these basic demands. What also ensued was the simultaneous centralization and decentralization of carceral space courtesy of digital computing networks.
Installing Computers in Cages
One of the ways that state authorities grappled with the carceral explosion was by centralizing administrative functions through computer technologies. Database management systems were embraced by prison administrators in efforts to expedite inmate processing. Before these databases, prison intake involved entering data on each incoming inmate’s aliases, birthdate, criminal record, fingerprints, medical/mental records, security threats, sex, and sexual offenses, among other factors. As the imprisoned population increased, screening each individual inmate became so burdensome that it led to overcrowded intake facilities. The state of affairs threatened to bottleneck the entire carceral machine. Inmate management databases were thus promoted by technocrats at the Justice Department, at the National Institute of Corrections, and by private firms, such as the SEARCH Group Inc., as remedies for these administrative challenges. As early as 1980, the National Consortium for Justice Information and Statistics began compiling directories for corrections management information systems. One of its first moves was to inventory the types of hardware, programming languages, and staff personnel (data entry operators, operations supervisors, programmers, systems analysts, terminal operators) in use at correctional facilities throughout the country (Figure 10). The consortium’s central goal was to create a knowledge community of technocrats and companies to brainstorm how technology could be put in the service of mass captivity. In 1982 in Illinois, the Criminal Justice Information Systems Division and Statistical Analysis Center initiated the Correctional Institution Management Information System Data Project to appraise the use of database management in the Cook County Department of Corrections. The project focused on finding computerized solutions to cleaning and circulating inmate data. Shortly thereafter, the Association of State Correctional Administrators, Bureau of Justice Statistics, Corrections Program Office, Federal Bureau of Prisons, and NIJ joined forces to design standardized templates for digital inmate processing. They proposed automation as the cheapest, fastest, and most effective way to slash the amount of time incoming prisoners spent in intake facilities. Carceral technocrats also developed classification rubrics to analyze these data to determine an inmate’s disciplinary, housing, and security requirements. Toward the end of the 1980s, the Office of Justice Programs and Bureau of Justice Statistics conducted a series of reports, such as Strategies for Improving Data Quality, to make data entry, data maintenance, and regulatory functions more efficient.
The slow creep of IT extended to one of the most totalitarian expressions of mass criminalization—the supermax prison. These hypersecuritized “prisons inside of prisons” were reserved for political dissidents and violent inmates. While the notion of indefinite administrative lockdown dates back to Alcatraz circa the Great Depression, it achieved broad normalization in the 1990s. Carceral power achieved its densest state in supermax facilities, thanks in part to tech firms. These facilities were designed with control rooms as nerve centers, where correctional officers could check each individual cell using radio frequency monitoring. Internal detection technologies including electrical circuit systems, light beam intrusion systems, motion sensors, and sound detection systems were installed to securitize ceiling crawl spaces, mechanical rooms, and ventilation openings. Computer-operated lock operating systems were put into cell doors so that they could be opened remotely. Telephones and intercom systems were fit with automated recording systems to listen in on conversations. Remote-controlled surveillance cameras were stationed to overlook elevators, living unit access, living unit day halls, pedestrian circulation corridors, and sally ports. These hyperregulated compounds were further fortified by automated perimeter security systems composed of electric field sensors, infrared sensors, seismic sensors, and video motion detectors. Wardens across the country championed computer-managed lighting and sensors as cost-effective substitutes for expensive watchtowers. Lethal fences capable of conducting up to ten thousand volts of electricity also emerged as common perimeter defense technologies at the edges of supermax apparatuses.
Prison administration could not escape some measure of technological displacement under these conditions. The RAND Corporation proclaimed that the mergence of computer technology and correctional management “changed the nature of prison functioning and of prison construction and design. The transition moved many prison functions from being personnel-intensive to technology-enabled.” In point of fact, as early as the mid-1980s, Denning Mobile Robotics Inc. began development of robotic prison guards. Standing three and a half feet tall and weighing two hundred pounds, the Denning Sentry was designed to detect intruders and smoke using Motorola 68000 microprocessors. The dream of digitally administered inmates was further pursued the following decade. One example involved using radio frequency identification tracking devices to conduct roll calls, notify staff when inmates were approaching restricted areas, and alert correctional officers when rival gang members closed in on one another. The NIJ and Space and Naval Warfare Systems Center sought ways to combine location-aware prisoner tracking, biometric devices, and CCTV surveillance to keep constant watch over inmate consumption and movement.
As the 1990s progressed, the IT sector was offering products and services for a myriad of prison processes, from access control to prisoner monitoring. For instance, the company GTL, which markets itself as the “corrections industry’s trusted, one-stop source for integrated technology solutions,” unveiled a database to provide “end-to-end prison management” throughout the cycle of sentencing, intake, cell assignment, meal planning, and release. New markets also emerged for a myriad of technologies to monitor staff and visitors. Backscatter imaging, electric field tomography, low-dose radiation body scanning, and orifice-scanning machines are some examples. By the mid-1990s, as the carceral population surpassed 2 million people, prison technologists, bureaucratic social scientists, and IT companies scrambled to understand how digital technology might best help manage the unwieldy scale of mass incarceration. Much of the development in this sphere was assisted by the National Law Enforcement and Corrections Technology Center (NLECTC), which grew out of the NIJ’s Office of Science and Technology in 1994. NLECTC’s main objectives included establishing equipment performance and technical standards, providing research and development support to the NIJ, and overseeing the transfer of information communication technology from the prototyping stage out into the correctional and law enforcement fields. A new market was born. Companies as big as Motorola and as small as Syscon took to designing software for corrections management, offender management, and jail management.
During NLECTC’s early days, federal and state-level prison administrators organized a country-wide report on correctional databases (Figure 11). The report gave birth to a standardized rubric of more than two hundred data elements, including information on inmate behavior, convictions, fees and fines, risk assessments, sentences, and socioeconomic profiles. Taking advantage of the growing carceral data environment, Federal Prison Industries, which oversees manufacturing in federal penitentiaries, developed a database that used Microsoft’s SQL Server 7 for data warehousing, electronic commerce, and enterprise resource planning to identify emergent security risks, prison transfers, and punishments. Correctional institutions in Colorado, Illinois, North Carolina, Pennsylvania, and Washington developed risk-assessment algorithms to predict inmates’ risk of recidivism, violations, and violence. These were intended to aid administrators in deciding cell assignments, security precautions, and disciplinary programs for inmates.
Carceral algorithms calculated inmates’ risk of violence and recidivism based on factors such as educational attainment, gang affiliations, language proficiency, mental health evaluations, socioeconomic background, and work history. They also calculated the relative significance of each factor for a number of possible scenarios (e.g., collusion or escape). Additional algorithms, such as the psychological assessment technology firm PAR Inc.’s Inmate Personality Assessment Inventory and the PAI Interpretive Report for Correctional Settings, were also developed during this period to assess the emotional and psychological states of inmates.
In addition to managing individual inmates, carceral technologists looked to digital computing to manage the arrangement of space. As the millennium neared its end, the American Probation and Parole Association convened in New York to explore how geographic information software could be implemented to organize carceral space, time, and subjects. The summit laid the groundwork for a geographic information database system called Correctional Mapping (CORMAP), which classified inmates by risk factor to assist prison administrators in determining cell and housing arrangements. CORMAP tracked the age, ethnoracial category, religion, and gang affiliation composition of each cell and housing unit. Some later versions of the software generated predictions of potential escape routes based on the architectural layout of the facility, group affiliation data, inmate locational data, and inmate interaction records. CORMAP also generated visualizations of the locations of gang members and sites of repeated disciplinary incidents. When linked to surveillance equipment (e.g., ankle bracelets, CCTV cameras), it could receive and map the locational data of an inmate’s every movement. NLECTC toyed with versions of CORMAP that kept logs of every person with whom an inmate has been in contact during incarceration. As the twenty-first century advanced, the carceral state’s technologists began to explore how prison space-time could be represented in three-dimensional models. During the same period, the California Department of Corrections and Rehabilitation (CDCR) joined forces with the LAPD to integrate geographic information databases into prisons. Police saw a great deal of potential for integrating digital mapping into prisons, as they were the “functional equivalent of precincts.”
The efforts on behalf of administrative technocrats and eager IT companies amounted to the dream of a truly totalitarian apparatus. Their imaginations ran wild with ideas to add biometric, disciplinary, family, financial, legal, locational, medical, and physiological information about inmates to the carceral state’s database infrastructure. They eventually summoned the image of a computerized cage in which all behaviors, movements, and risks were quantified and analyzed in real time. Early into the new millennium, the NIJ and Space and Naval Warfare System Center began to investigate how biometric technologies could be implemented in correctional facilities to manage access to different areas. Fingerprint and iris scan technology would regulate the movements of inmates, prison officers, and visitors. Florida State University’s Center for Criminology and Public Policy collaborated with the state’s Department of Corrections to optimize its Correctional Operations Trend Analysis System (COTAS), which analyzed both historical and real-time data concerning individual inmates, institutional performativity, and prison incidents. The system, whose origins lie in the state’s Bureau of Research and Data Analysis in the late 1990s, made a series of predictions by probing a multiplicity of databases, including the Fast Access Secure Tracking system (for inmate visits and volunteer activity), the Human Resources Database, the Management Information Notes System, the Offender Based Information System, the Inmate Gang Database, and the Use of Force Database. One of the most celebrated features of COTAS was its purported ability to determine an inmate’s threat level based on age, bed category, drug tests, gender, involvement in violent events, number of violent events, time served, and racial identity (which only included “white” and “nonwhite”). NLECTC and the U.S. Energy Department’s Savannah River Technology Center prototyped more advanced versions of similar software that provided real-time, multilevel, three-dimensional simulations of prison facilities and the people inside, replete with data on parolees, prison facilities, prison officers, probationers, and inmates. Similar systems were installed throughout prison facilities for information on booking, cell checks, inmate movement, medical records, and visitations. Some companies, such as Kiosk Information Systems, Keefe Group, SeePoint, SmartJailMail, and TurnKeyCorrections, dove into the automated kiosk industry to grant inmates access to email, conduct commissary transactions, participate in video visitations, and schedule appointments, sometimes for a fee.
Marx’s analysis of large-scale industry helps us to understand this other Satanic Mill, the computerized prison. Indeed, the promise of digital technology companies was a prison managed like an orderly factory. The continuous automation of mechanisms that captured the criminalized subject, drove it from point to point through the criminal justice system, and regulated its movements and distributions through space was supposed to give way to a “vast automaton” of great efficiency and scale. On the micro level, computerization was intended to help optimize the prison’s architectural, functional, and hierarchical political anatomy. Information systems were intended to assist carceral administrators in distributing devalued humans through cells and corridors; organizing carceral enclosures according to sociological and penological categories; enhancing abilities to supervise, immobilize, and capture inmates; and ranking inmates and prison personnel according to normative baselines pertaining to race, educational background, employment history, and numerous other attributes (Figure 12). What is more, the technologies were said to enact a type of time discipline similar to the factory through automated lighting, roll calls, and daily regimens determined by risk-assessment tools. But the metaphor stops here, and racial difference is the reason. The computerization of the carceral apparatus was not driven by extracting surplus from production processes or manufacturing docile consumer-workers; rather, it was driven by cities determined to quarantine economically devalued populations and by information capital seeking new profit frontiers.
In addition to provoking new ideas about establishing total control inside correctional facilities, digital technology has facilitated the diffusion of carceral power over public space. Prison development siting has become increasingly influenced by computer software over the past two decades. Near the turn of the century, planning agencies such as California’s Office of Planning and Research started to design location-siting algorithms to help determine where to construct new halfway houses, jails, and prisons. The algorithms identified construction sites based on a series of factors, including land uses, infrastructures, and transport connections. For existing prisons, municipal planning agencies also used management applications to monitor the conditions of correctional land holdings, buffer zones, and wastewater fields.
Digital computing has also come to play a role in determining where inmates are released once their sentences terminate. As the Great Recession placed increased strain on public expenditures, some state officials turned to risk algorithms to defray the mammoth costs of mass incarceration. Algorithms designed to assess inmates’ likelihood of recidivism were embraced by courts, prisons, and juvenile detention centers to help determine which inmates ought to be eligible for early release or work programs. To be certain, risk assessments have been applied in social service apparatuses since the 1920s. In the 1970s, risk scores were used to predict future behaviors of juvenile offenders. Two decades later, the Office of Juvenile Justice and Delinquency Prevention released its Comprehensive Strategy for Serious, Violent, and Chronic Juvenile Offenders, which expanded the practice to identify low-risk inmates and relocate them into parole or probation programs. But the epic scale of the carceral crisis saw to it that these algorithms were computerized. University programs like the Center for Evidence-Based Corrections and public agencies like the New York City Criminal Justice Agency worked diligently to codify risk instruments. These were designed to help administrators determine release eligibility based on “static values,” such as gender, mental/health problems, and prior record, and “dynamic values,” such as age, education level, employment, and marital status, in languages amenable to computer programming. Developed in conjunction with criminologists, these algorithms typically coded eight variables as the most powerful predictors of criminal tendencies: antisocial associates (association with offenders), antisocial personality pattern (adventurous, impulsive, pleasure-seeking behaviors), antisocial cognition (negative attitudes toward legal authority, rationalization of criminal activity), family/marital status (estrangement from parents or spouse), history of antisocial behavior (arrested at young age, large number of prior offenses), leisure/recreation (low participatory levels in recreational activities), substance abuse (alcohol or drug abuse, save tobacco), and school/work (low participation and/or achievement in educational institutions or labor markets). Companies such as the Northpointe Institute Orbis Partners and Multi-Health Systems turned these instruments into software packages that quickly spread across the country.
The Herculean efforts to make the carceral state digital went hand in hand with the extension and thickening of the parole and probation apparatus. This was due in no small part to the fact that between 1980 and 2015, the probationer and parolee populations increased fourfold and fivefold, respectively. As a consequence, a variety of geographic information-based systems crafted to manage parolees and probationers throughout public space began to appear on the market. In 2001, the New York City Department of Probation (NYCDOP) unveiled the Statistical Tracking and Reporting System (STARS), which employed geospatial software to help probation offices manage information on sixty thousand adult probationers and twenty-five thousand juveniles. City officials boasted that STARS generated forty thousand pre-sentence investigations for the U.S. Supreme Court and criminal courts and seven thousand investigations and recommendations for family courts. By 2008, NYCDOP had adopted geographic information systems to produce digital maps of recent arrests, gun stores, drug trafficking areas, liquor stores, and other probationers to assist parole/probation boards in determining where probationers should be released and were authorized to travel. To coordinate the geographies of correctional supervision, software company MapInfo created an application to optimize the distribution of parole and probation officers throughout city space. The company adapted an older application to generate districts based on caseload distributions of parolees and probationers. The districts were meant to help administrators select caseloads for probation/parole officers (PPOs) and determine home visitation assignments. In a similar vein, Esri repackaged older software packages to determine route sequences for PPOs during home visits. Software applications like these were designed to minimize the amount of time it took for PPOs to establish contact with parolees and probationers. The overriding objective was to establish a logistical correctional apparatus whose capillaries spread across urban space.
Technological displacement came to loom over PPOs under such conditions. Companies like DynaTouch, Napo, and Northwest Ohio Regional Information System have taken to producing automated kiosk terminals exclusively for probationers to report themselves. Usually located in courthouses, police precincts, and probation offices, the terminals allow parolees and probationers to check in via biometric identification and then answer questions pertaining to drug tests, employment, and housing. Technology firms, including AnyTrax and Fieldware LLC, offer automated telephone- and internet-based reporting for low-risk subjects living in remote areas. Some of these remote reporting terminals come equipped with a breathalyzer, polygraph, voice stress analyzer, and voice recognition technology.
The most significant development in the digital extension of correctional supervision has been the use of the Global Positioning System in parole and probation management. This location-aware form of correctional supervision has quietly advanced the correctional management of everyday life. Early forms of electronic correctional supervision using radio frequency technology first appeared in the 1960s in Cambridge and Boston, Massachusetts. Dubbed “electronic parole,” the program involved parolees wearing portable transceivers. A researcher from Harvard’s Science Committee on Psychological Experimentation who assisted with the technical side of the procedure saw the radical implications of this new dimension of supervision. “When specific offending behaviors,” he asserted, “can be accurately predicted and/or controlled within the offender’s own environment, incarceration will no longer be necessary as a means of controlling behavior and protecting society.” It would be some time before electronic monitoring was truly unleashed. In 1983, a district state court in New Mexico ordered multiple parole violators to be supervised through GPS tracking. This revival of electronic supervision was also stimulated by Reagan’s Military Cooperation with Civilian Law Enforcement Act (1981), which allowed military technology, such as GPS devices, to be transferred to municipal law enforcement. Not only did the act contribute to the militarization of racial criminalization but it also extended its field through electronic correctional supervision.
Three catalysts would expand electronic supervision even further, moving further into the present century. The first was organized resistance against mass imprisonment. Public awareness for prison reduction gained national traction through the formation of the Sentencing Project in 1986 by Malcolm C. Young and of Critical Resistance by Angela Y. Davis, Ruth Wilson Gilmore, and Rose Braz in 1997. Over the next decades, momentum from these struggles carried over into mainstream political discourse. This was due in no small part to the appearance of Michelle Alexander’s influential book The New Jim Crow and the Brown v. Plata (2011) U.S. Supreme Court ruling mandating the observance of population limits in prisons the following year. Moreover, a 2015 federal report exposing the excessive force, rape, and solitary confinement at Riker’s Island prison contributed additional fuel to resistances against the carceral state. To pacify mounting social forces, federal, state, and municipal authorities across the country slowly began to recommend electronic supervision as a humane alternative to imprisonment. The second catalyst in the spread of electronic supervision was the Great Recession and the inertial force of austerity politics that followed. As state budgets across the country dried up, line items were scrutinized with extraordinary care. In 2010, the average annual cost of imprisonment was revealed to be more than $31,000, a figure that exceeded the labor market value of inmates. The Vera Institute reported that states spent more than $43 billion on prison expenditures in 2015 alone. Electronic supervision was touted as one means of cutting these expenses, as the total number of probationers rose to quadruple of it what it was when the Reagan administration commenced. The third catalyst was information capitalists, who, with the cleverness of a parasite, seized the opportunity to profit from decarceration. For instance, the Center for Media Justice discovered that ankle bracelet suppliers Attenti, BI Incorporated, Satellite Track of People, and Sentinel Offender Services alone generated more than a billion dollars of revenue on an annual basis.
Electronic supervision inscribes the correctional apparatus deeper into urban space and interiorizes it in the lifeworlds of criminalized populations. Upon enrollment in the electronic monitoring apparatus, one is given a digital shackle to wear that radiates locational data to PPOs, nearby police patrols, and, in some instances, real-time crime datacenters (see Figure 13). The device is meant to ensure that one is at an authorized location, which Africana studies scholar James Kilgore has shown usually involves house arrest. Should the monitored subject enter unauthorized space or tamper with the device, an automated alarm is sent to the parole/probation office, nearby police patrols, and victims in cases involving domestic and sexual violence. Digital supervision programs like these were pioneered in California. Early in the 2000s, the CDCR used risk-assessment tools to identify the most violent people registered in its gang database. The individuals were then tagged with GPS shackles and told where in the city they were (dis)allowed to go. The shackles were also used to generate data to help authorities map gang turf more accurately. For instance, the GeoShadow initiative of the Oklahoma Department of Corrections, NLECTC, Office of Science and Technology, and University of Oklahoma’s Center for Spatial Analysis used similar methods to track high-risk subjects. With location-aware supervision, mobility, not visibility, is the trap. In some instances, people enrolled in these programs are permitted travel to specified locations for specified amounts of time. In some cases, monitored subjects must attend rehabilitation centers, school, psychological evaluations, or workplaces. In many cases involving juveniles, data produced from the monitored subject’s drug tests, home visits, school/work reports, and psychiatric session are also fed into software programs that scan for fluctuations in risk ratings. In these instances, the technology helps the correctional apparatus modulate rehabilitation centers, schools, and workplaces into disciplinary satellites. Electronic supervision cases involving domestic, gang, and sexual violence often involve establishing exclusion zones, or electronic barriers that trigger alarms when breached. Exclusion zones bisect urban space into areas in which parolees or probationers may and may not travel and thus form an electronic grid for carceral urbanization. Disciplinary society and control society collapse in such a scenario. The digitized expression of carceral space does not replace discipline with control; it cross-fertilizes the two.
It would not be unreasonable to see electronic supervision as a new horizon of carceral statecraft. This apparatus that combines uninterrupted monitoring and evaluation invokes images of the Public Safety Bureau in Gen Urobuchi’s manga Psycho-Pass. In it, the bureau uses proprietary technology to monitor the physiological states of citizens for deviant vectors. Urobuchi’s dystopian vision was partially realized in juvenile electronic supervision programs in New York’s Division of Probation and Correctional Alternatives. Some of these programs required juveniles to regularly submit “status of life” evaluations that provided information about drug tests; living situations; time spent in job training, school, or work; and psychiatric diagnoses. The information was entered into digital profiles that were analyzed for emerging risk. Here we see a glimpse of a type of carceral power whose audacity is only matched in Urobuchi’s fictional account. One mistake, and you are forced into a cage. In fact, in 2014, more than one-quarter of all prison admissions were for technical violations of parole or probation, such as associating with known offenders, failing a drug or alcohol test, failing to report to a parole/probation officer, missing appointments with social service agencies, or violating geographic restrictions. Unlike Urobuchi’s account, however, electronic supervision is not applied universally throughout the body politic. It is instead reserved for tracking the economically devalued and racially stigmatized populations who dwell in the cracks of urban centers of global capital.
By the mid-2000s, an estimated one hundred thousand parolees, probationers, registered gang members, sex offenders, and work release program participants were under GPS-based supervision. This represented a ninefold increase in the number of electronically supervised subjects in just over a decade. Cases involving domestic violence, drunk driving, undocumented immigration, sex offenses, truancy, and a range of minor offenses including ordinances violations and trespassing are now punishable by digital detention. Increasingly, the electronic bracelets are equipped with blood alcohol detectors and breathalyzers. Academics eager to get a piece of the industry have already explored ways of equipping the bracelets with conducted energy devices capable of electrocuting monitored individuals who violate parole/probation conditions. In 2006, California Proposition 83, also known as Jessica’s Law, mandated that sex offenders be placed under GPS supervision for life. Four years later, thirty-three states had begun to employ electronic supervision for sex offender tracking. Electronic monitoring has also been adopted by New York City’s Immigration and Customs Enforcement branch for its alternative detention initiative. Used primarily on undocumented immigrants from Central America, the initiative is part of a nationwide program used on more than forty thousand criminalized immigrants. In addition to GPS devices, the program uses voice recognition software so that employers may verify a monitored immigrant’s presence at workplaces. Here digital networks are mobilized as a conduit of border relations and something resembling privatized prisoner labor.
The inefficiency of the court system during the apex of the War on Drugs was prodigious. But this state of affairs was propitious for software companies. The utopian vision of a court apparatus that operated with clocklike precision peddled by these companies was captured by the Illinois Criminal Justice Information Authority in 1993:
A criminal-court judge sitting in her chambers mulls over a difficult fraud case. She picks up a pen-like device attached to a computer to select a case from an on-screen list. The computer displays videotaped trial testimony with captioned text. Using the computer pen, the judge writes a key phrase onto the screen and the computer fast-forwards to that portion of the testimony. After reviewing the video passage, the judge switches to a criminal records database before calling up an artificial-intelligence-based software program to determine the appropriate punishment.
artificial intelligence, in which a computer replicates the knowledge and reasoning of a skilled professional, may someday assist judges in sentencing or attorneys in selecting jurors. While development of those applications is just beginning, a number of courts are already using a sophisticated computer program to evaluate the likelihood of substance abuser rehabilitation.
Though digital technology has not penetrated courts to the same extent as the other branches of the criminal justice apparatus, it is still important, because courts play a vital part in determining the size, profile, and, in some instances, geography of criminalized populations. The rapidity with which the carceral state expanded during Reagan’s War on Drugs in the 1980s placed heavy strain on criminal processing procedures. Human Rights Watch reported that the decade saw drug arrests increase from 376,155 to 981,381, about 160 percent. And as the War on Drugs raged, it became all the more important for courts to pronounce judgments and punishments swiftly and to distribute those whom it convicted between community supervision, house arrest, and jail in an efficient manner. If the court was not up to this task, many bureaucrats worried, the carceral colossus would buckle under the pressure of its own weight.
Originally, the computerization of judicial power revolved around digitizing documents to reduce trial time and provide judges access to the internet to cross-verify references and participate in remote deliberations. It was, however, the application of database management systems to criminal proceedings that judicial technocrats raved about the most. Prosecutors in Los Angeles, New York City, and St. Louis were using computer-aided transcript systems already in the 1970s and 1980s. These systems, originally developed through Law Enforcement Assistance Administration (LEAA) funding, were designed to manage the deluge of cases following the launch of the War on Crime. The uptick of political prisoners during this period was making it all but impossible for attorneys to process paperwork in an accurate and timely fashion. This resulted in backlogs of pretrial defendants, which threatened to overload pretrial detention facilities. Database management systems were posed by IT companies and technocrats as remedies. Justice management databases were rolled out with the promise of ushering in the age of the paperless courtroom. By digitizing records, companies such as ICON Software Corporation and Tyler Technologies marveled, courts would also be able to convert storage room and cut back on human clerks. Several technologies were also developed to hasten procedures. Early in the 1970s, the LEAA teamed with Inslaw Inc., a technology corporation that specialized in administrative database management systems, to develop prosecutorial software. The collaboration gave birth to the Prosecutor Management Information System (PROMIS), which offered dozens of functions for legal proceedings. PROMIS was designed to administer data about court costs, court orders, courtroom scheduling, decisions, defendants, evidence, fines, jurors, offenses, plaintiffs, and restitutions. It also allowed police to transmit information about arrestees, victims, and witnesses to prosecutors and administrative agencies. Furthermore, PROMIS automated the assignment of cases to attorneys based on experience and workload. In New York, the Information Systems Department of Criminal Justice Agency Inc.—a legal service consultant founded by a New York industrialist—introduced a database that stored case-processing data from courts, demographic data about communities, and incident data from the police.
The drug war pushed forward judicial computerization. By the 1990s, the U.S. Attorney’s Office had unveiled its own customized database, which organized data on cases, defendants, investigations, gang warrants, general crimes units, narcotics, and wiretaps. It also had geographic information functionality so that attorneys could conduct queries by location and view spatial distributions of their dockets. These specs were meant to bolster information sharing throughout the administrative state apparatus. Videoconferencing also emerged in courtrooms to accelerate offender processing. Video technology first emerged in an Illinois court in 1972 for a bail hearing and came to be used in the majority of courts for various ends. Use of the technology expanded significantly during the 1990s to alleviate pretrial overcrowding, schedule arraignments for people arrested during courts’ nonoperating hours, and reduce costs of moving inmates from jails to courts. Inside courtrooms, video technology was instituted so that alleged victims could offer real-time testimonies from remote locations. Touchscreen technology was also introduced to witness boxes for witnesses to look at images and hear audio of defendants and suspects. Document scanners, email, real-time court transcripts displayed on courtroom monitors, two-way video, and wide area networks have also reshaped the materiality of the judicial apparatus. By the early 2000s, over 95 percent of federal district courts used audio conferencing equipment, evidence cameras, and internet connections. More than 80 percent used bench monitors, witness stand monitors, and video conferencing equipment. In some cities, judges digitally sign orders, and trial lawyers use trial presentation applications on laptops and tablets to organize designation reports, deposition videos, exhibits, deposition transcripts, and witness/trial notebooks.
What began to emerge from the application of digital computing to the judicial apparatus was the dream of a logistical complex that processed convicted populations without interruption. This was bolstered by endeavors to digitize the transmission of evidential data to judges, jurors, and legal teams. In the late 1990s, the National Task Force on Court Automation and Integration conjured the specter of a frictionless criminal processing environment where an arraignment judge sitting behind a computer could pronounce judgment within minutes of a suspect’s arrest. The task force imagined a day when alimony orders, criminal history records, drug test results, fingerprints, gun purchase applications, protective orders, sexual offense registration status, and warrants would be instantly retrievable by judges. During booking, automated databases would assemble data on defendants, case characteristics, charges, sentencing, witnesses, and victims for attorneys, judges, juries, and prosecutors. These databases would be equipped with algorithms to determine a defendant’s harm to society, risk of flight, and risk of future crime to help judges and prosecutors make pretrial preparations. Risk-assessment tools were also proposed by judges for various decision-making tasks concerning bail, bond amounts, diversion programs, guilt/innocence, rehabilitation requirements, the mental health of accused offenders, releases on recognizance, and sentencing. The Conference of Chief Justices, Conference of State Court Administrators, and National Center for State Courts all lobbied for making risk-assessment tools part of standard procedure. Their ultimate stated goal was to speed up court appearances to the point that they could take place within twenty-four hours of booking.
Dozens upon dozens of corporations, including eFORCE, ExhibitView, Moonlighting Software, Synergy International Systems, and Tyler Technologies, have rushed to corner market shares in judicial database systems. Near the end of the 1990s, the Center for Applied Studies of the Environment, the Southern District of New York, and the U.S. Attorney’s Office teamed up to create a case-tracking and mapping database system explicitly for the narcotics trade. Similar database technology is now marketed as a way to manage arrest reporting, court docketing, jail booking, mobile dispatching, and offender registries. At the municipal level, which deals primarily with minor adjudications, court management systems are used for perfunctory tasks of collections, docketing, and scheduling. Appellate and juvenile courts use specialized software that has bank account applications for collecting and distributing childcare, fees, fines, garnishments, restitutions, and credits for offender work programs.
It is perhaps impossible to measure the effect that these technologies have had in adjudication. In many cases, access to the algorithms is blocked from the general public. Public safety is invoked when the software is developed by public–private partnerships; intellectual property is invoked when it is developed privately. What remains clear, though, are the social categories that these algorithms are programmed to criminalize. Most propriety software generates risk scores according to the “central eight” commonly cited factors in criminal psychology. These search inmate data for high levels of antisocial associates, antisocial cognition (attitudes, values), antisocial personality (stimulation, low self-control), criminal history, family problems, employment instability, low participation in social leisure activities, and substance abuse. Others process data from prisoner surveys and personal records, including data on depression, educational achievement, length of unemployment, powerlessness, psychological handicaps, reliance on welfare, resentment, and minority status. These predictive tools do not explain criminality; rather, they describe those whom the carceral apparatus is literally programmed to micromanage. They describe the age, race, gender, psychological profile, and socioeconomic status of people targeted for more than a half century by the apparatus of mass criminalization. Perhaps the most insidious thing about the scientific veil is that it conceals the barbarism of the criminal justice system that it services. These digital maps, machine learning algorithms, and statistical curves are developed in service of state apparatuses that round up millions of people at gunpoint and lock them in cages; that encage young people for years at a time for selling marijuana; that are characterized by routine sexual assault; that execute people using experimental concoctions of lethal drugs; and that murder twelve-year-olds in plain sight and on camera with impunity.