How Technology-Assisted Surveillance Promotes Racial Bias in Policing

Policing and incarceration in the twenty-first century are not wholly human practices. The strategies, tactics, and encounters that organize the carceral state are developed and prosecuted in collaboration with inhuman technologies that “we” do not and cannot fully understand. In fact, the demotion of the human agent in policing is often by design. Making sense of carceral power at this moment requires that critics approach policing the way we do other industries that are widely acknowledged to be fully captured by digital media, like finance.

Policing is, now, an inhuman practice in both an ethical and a technical sense. People might set policy and parameters for what we’d like our carceral institutions to do. But, practically speaking, none of it is imaginable or executable without digital technologies that are fundamentally black-boxed, inscrutable to individual cops on the beat and institutional decision-makers alike. This, as much as anything, is the key point of the carceral surround: “We” humans are the products of the twenty-first-century carceral technology industry as much as its authors.

I’ve developed the carceral surround to name the logic that connects the NYPD’s 2020 siege of Black Lives Matter activist Derrick Ingram, who was wanted for yelling into a bullhorn during that summer’s protests, to police agencies’ precision targeting and abduction of international students opposed to genocide in Gaza, to the surveillant mundane of everyday life under what Jackie Wang names “carceral capitalism.” This logic is extensive and immanent to the social world. It organizes the network of computational technologies that collectively work to instantiate carceral power as a bedrock and reality-defining force of contemporary capitalism. This force flows from a populational logic that uses statistics and algorithms to distribute the state’s violence work.

Instead of law enforcement, policing shifts to violent secure against mathematical possibility by acting against living humans now.

In the 2010s, when the massive generation and capture of data became simply referred to as “big,” highly capitalized information technology companies promised that they and only they could make sense of the new hyperobject. This was as true in the carceral field as it was in other industries. Silicon Valley insisted that their tools could finally eliminate uncertainty and subjectivity from juridical decision and instantiate a perfectly mathematical official antiracism in criminal justice. The momentum in many criminal legal enterprises began hurtling toward digitally automating this analytic regime of knowledge.

Nowhere was this convergence more spectacular than in California, which launched an unprecedented Supreme Court-mandated effort to depopulate its “Golden Gulag” of state prisons at the beginning of the decade. Microsoft, the carceral state’s most important IT partner, wanted in. California’s Proposition 47 reduced certain drug convictions from felonies to misdemeanors and allowed inmates and formerly incarcerated people to petition the court for resentencing. It didn’t permit re-sentencing to those with any prior convictions for “sex offenses including rape, child molestation and other sex offenses; murder, attempted murder and solicitation to commit murder; assault with a machine gun on an officer; or any serious or violent crime punishable by a life sentence or death.”

nBut it compelled the resentencing of any other inmates “unless the court finds an unreasonable risk to public safety,” determined by “the offender’s criminal history, the types of crimes committed and when they occurred, the extent of injury to victims, the length of prior prison commitments, the inmate’s disciplinary and rehabilitation records while incarcerated, and any other relevant evidence.”

The passage of Prop 47 meant that the criminal legal system had slightly less leeway to lock “modestly educated women and men in the prime of their lives” in cages. But because of the efforts of officials like Commander Will Brown, the San Diego County Sheriff’s Department officer in charge of its Prop 47 transition, it became an opportunity to imagine a redesign of the entire criminal legal system. Brown seized on the law as an opportunity to use risk to “basically level the playing field, re-evaluate how criminal justice happens in our county and remove all the labels that we’ve learned over the past three years: split sentences, straight sentences, post-release community supervision, 999 [the term for someone sentenced under California’s three strikes laws]…all those labels I think should go away,” he said.

In their stead, Brown envisioned “risk and risk alone” determining “how we spend resources and how we provide for public safety.” If the police and the courts determine that someone is at a high risk of committing new crimes, he argued that the police owe it to their “elected constituents to reduce that risk by spending resources on an offender to lower their recidivism so they’re not creating new victims in the community.”

The sheriff’s department turned to Microsoft to help accommodate this vision. The company would link the different agencies’ data silos, provide software and hardware to facilitate the GPS tracking of paroled people convicted of crimes, and develop an efficient data analytics system to analyze in real time agency databases and surveillance information to determine risk levels of persons that the police encounter. This profoundly racialized idea of security effectively transforms living people into possible future threat. Instead of law enforcement, policing shifts to violent secure against mathematical possibility by acting against living humans now.

Ruha Benjamin calls the collection of tools that are designed to produce this racial power “the New Jim Code,” and population racism works to think through how this New Jim Code works to produce new forms of value in a 21st century political economy. As terrifying as so many of the technologies that constitute the carceral surround are (or purport to be), the reality is that in practice many of them work terribly. Despite truly astronomical levels of funding, most of the flashiest and most future-shocked crime-control technologies fail and fail consistently. ShotSpotter, which is designed to automatically detect and pinpoint gunshot locations, has been found to cost NYC tens of millions in wasted funds as it consistently directs officers to incorrect locations. Commercial predictive policing technologies (which have largely been brought under ShotSpotter’s parent company SoundThinking’s umbrella) do little more than ratchet up racist overpolicing and offer bad predictions. Automated facial recognition software is notoriously inaccurate, and especially so against people of color, who are most likely to be targeted by police.

Even more seemingly banal technologies like body-worn cameras have had no discernible impact on police behavior or on departmental transparency. Efforts to integrate these cameras with software that would provide real-time automatic threat assessments to officers (by analyzing the body language of whomever police are interacting with) have gone nowhere. Likewise, NYPD-issued smartphones were imagined to fully integrate beat cops with the city’s automated surveillance, command, and control apparatus, but the main impact has been to create an indelible image of waste and indifference: of officers standing around staring at their phones all day. This list could go on.

And yet research and investment continue to pour into digital policing technologies even as the failures mount. In certain spaces, like community supervision (probation and parole), automated surveillance tech is now de rigueur, even as its brittleness consistently remands people to cages over sometimes literally human-imperceptible infractions. What James Kilgore calls “e-carceration” continues to proliferate as an alternative to lockup as overburdened cities and states seek the supposed benefits of mass incarceration’s spatial fix without incurring its attendant fiscal cost. Meanwhile, a cowed political class meets every crisis caused by the monumental violence of policing and prison with calls to direct more funds to the instigators of racial-sexual oppression and to more heavily invest in technologies and training. In other words, repeated failure catalyzes an ever-deeper investment in the carceral surround as a horizon of governance.

For Commander Brown, public safety is determined by risk that technocrats meet behind a closed door to determine. In fact, Brown argues that the principal advantage that Microsoft’s package offers to “criminal justice partners”—the term the company uses for all players in the public-private partnership of policing—is that it facilitates the smooth and seamless communication between agencies necessary for them to “share resources…[and] information to make your county safer” without making such communications public. Rather than taking place among elected or public officials, these conversations take place among the “people below that position” of representatives and public-facing managers who can be “very frank and can just put it on the table in a way that your elected and chiefs can’t.”

Digital technology captures and shapes this violence as both best practice and as a model for society. It not only makes such an order of things “normal,” it makes it seem “natural” or, worse, “right.”

Let’s not let this moment pass. What about how police agencies understand criminality and risk can only be put on the table and frankly stated behind closed doors? What important insights into policing can’t elected officials or chiefs say out loud? To say that the answer is race is both obvious and not enough. More specifically, it is population racism and the animating power of statistically racialized threat to do speculated harm. This negatively racialized fear of future harm is always posited against an equally racialized bearer of property that is harmed. It is, in other words, an imagined Blackened harm that animates these technocratic assessments of risk and precisely the whiteness of the harmed that must “very frank[ly]” remain (un)spoken only behind closed doors. This is the very logic of police power.

Prisoners are created through policing, which has been defined for decades by “zero tolerance,” “order maintenance,” “community policing,” or “broken windows.” There are some differences in each of these approaches, but the overriding philosophy that unites them is that it is the job of police to determine what counts as a “community” and then to maintain community “order.” Cops do this with the tools available to them: They criminalize and punish aspects of daily life that do not comport with a department’s sense of what order means. Police deploy a host of tactics that target “disordered” bodies in “ordered” spaces and “ordered” bodies in “disordered” spaces. This entails criminalizing everyday life and everyday forms of community. “Order” maintenance proceeds by disordering social bonds and community ties. A sidewalk party to celebrate a six-year-old’s birthday? Cleared by cops. A neighborhood cookout? Not without police permission. Black and Brown teenagers hanging out on the corner of their block? Inherently dangerous. Digital technology captures and shapes this violence as both best practice and as a model for society. It not only makes such an order of things “normal,” it makes it seem “natural” or, worse, “right.”

The only position capable of contesting the expanding power and violence of the carceral surround is abolition. To quote Ruth Wilson Gilmore, “Abolition requires that we change one thing, which is everything.” By offering the necessity of changing “one thing, which is everything” as its opening gesture, abolitionist thought breaks a cultivated feedback loop in which the insistence that carceral power is an isolatable part of the social world works to reinforce its place at this world’s center. Abolition, in this reading, is broad. It encompasses, of course, the elimination of police and prisons. But also, to quote Stefano Harney and Fred Moten, “the abolition of a society that could have prisons, that could have slavery, that could have the wage, and therefore not abolition as the elimination of anything but abolition as the founding of a new society.” In refusing and critiquing analyses that ask after how carceral technology is used and how it can be improved, abolition opens up the space to genuinely think critically—to ask not only what these technologies do but what can be built in their place.

_______________________________

Adapted from The Carceral Surround by R. Joshua Scannell, published by the University of Minnesota Press. Copyright © 2026 by the Regents of the University of Minnesota. All rights reserved. Used by permission.