Surveillance, Policing, and Algorithms

Since the brutal murder of George Floyd by the Minneapolis Police at the onset of last summer, there has been a resurgence of political energy amongst the American population centered on significantly reforming policing in the United States. Ranging from demilitarization to defunding, there has been no shortage of policy proposals issued by an endless assortment of governmental bodies, academic institutions, think tanks, non-profits, etc., as a means to promote greater accountability amongst a public institution that has fully exerted its monopoly on violence upon the American people. A critical area of policing that is in desperate need of stringent regulation is the seemingly unfettered use of surveillance tools. Specifically, surveillance tools that are embedded with artificial intelligence and machine-learning, algorithmic functioning. Some of these tools include but are not limited to: predictive policing software, facial recognition, automated license plate readers (ALPRs), and risk assessment scoring.

The utility of these kinds of surveillance technologies is seen in a few ways. First, it allows for a mass extraction of data points from various sources. Second, it allows for automated processes of analyses such as pattern recognition and data point linkage and connection. Third, based on the machine-learning of patterns and connections, it allows for predictive analysis and the alerting and signaling to end users (police) of the technology of potential risks or threats. These tools have be used to inform law enforcement which people and places to monitor. Moreover, they have assisted law enforcement in determining which people are potential threats to public safety and order. Consequently, people will have interactions with police, will be arrested, will be charged, and will be imprisoned based on the application of these surveillance technologies.

To this point it may be unclear as to what the problem is in regard to these police using these surveillance technologies. Surely, law enforcement is in need of some tools in order to root out and investigate crime. But before we fully address the problem with these tools we need further background regarding the functionality and utility of policing. To start, it must be acknowledged that the presence of an elite grouping of people, or a ruling class, who hold a disproportionate amount of political and socioeconomic power relative to the rest of the population has been a constant theme since the inception of the nation. Institutions and structures (of the government and economy) that form the backbone of the American state have been historically designed by and for the exclusive benefit of certain individuals: namely white, land-owning, generally wealthy males.

One of the most prominent of these institutions is law and the criminal justice system, in relating these institutions to that of the powerful, sociologist Richard Quinney argues in his book The Social Reality of Crime that, “Although law is supposed to protect all [residents], it starts as a tool of the dominant class and ends by maintaining the dominance of that class. Law serves the powerful over the weak…Yet we are all bound by that law, and we are indoctrinated with the myth that it is our law.” Consequently, there is a robust history of non-privileged classes and groups in American society challenging this unequal distribution of power through various resistance methods. Examples of these challenges include, but are not limited to: the abolitionist movement, organized labor movements, women’s rights movement, civil rights movements, anti-war movements, and most recently the resurgence of criminal justice reform and anti-police brutality movements. Unsurprisingly, all of these challenges have been met with immense pushback from the ruling class, expressed through government, i.e. the police, due to the significant threat they impose on their source of power. Sociologist Alex Vitale, thoroughly documents these episodes throughout his book The End of Policing and perhaps one the best takeaway points from this work is this: “The myth of policing in a liberal democracy is that the police exist to prevent political activity that crosses the line into criminal activity, such as property destruction and violence. But they have always focused on detecting and disrupting movements that threaten the economic and political status quo, regardless of the presence of criminality.”

The authority of the government is vested into the police to enforce the law, which itself is typically crafted to meet the interest of the ruling class. Thus the police serve as an extension, or more fittingly, a weapon and a shield of this power dynamic. Each iteration of the police has been imbued with the authority to use force as both a weapon of repression and as a shield to protect power and privilege from challenges and threats. As a result certain techniques and tools are used by the police to counter and neutralize these threats. One such tool is surveillance technology.

Perhaps one of the most succinct passages that captures the essence of this idea comes from a report entitled “Before the Bullet Hits the Body: Dismantling Predictive Policing in Los Angeles,” authored by the Stop LAPD Spying Coalition wherein they state, “Communities of color, immigrants and the economically marginalized are the primary targets of these modes of surveillance… It is yet another tool, another practice built upon the long lineage including slave patrols, lantern laws, Jim Crow, Red Squads, war on drugs, war on crime, war on gangs, war on terror, Operation Hammer, SWAT, aerial patrols, Weed and Seed, stop and frisk, gang injunctions, broken windows, and Suspicious Activity Reporting (SAR).”

Finally, in contemporary society it is not a secret (especially since the Snowden revelations) that surveillance is ubiquitous in the US. It is ostensibly utilized for the promotion of national security and public safety and there is a truth to this claim. Yet, in a society that remains highly unequal with the existence of an incredibly privileged ruling class it remains relevant to point out, as privacy scholar Jeffrey Vagle argues in his article “Surveillance is still about power”, “…surveillance is, at its core, about the establishment, use, and maintenance of power…even the most common surveillance practices have a power dynamic that too often shifts from generally beneficial to abusive.” In sum, surveillance is an expression of power. It is also a tool wielded by the institution of law enforcement, itself an arm of the ruling class. Surveillance mechanisms are designed and have been utilized to preempt organized dissent from the status quo: which is the highly unequal distribution of political and socioeconomic power.

With this political and historical context in mind, we can now return to the issue at hand: police use of surveillance technologies embedded with artificial intelligence and machine-learning, algorithmic functioning. Rather than outlining each of the surveillance technologies listed in the introduction in regard to their various features and components, I will summarize some of the major concerns that have been put forward in the literature surrounding this topic. Sociologist Sarah Brayne has shown in her work, most notably her article entitled “Big Data Surveillance: The Case for Policing” that there are noteworthy implications for the reproduction of inequality through the utilization of these technologies. For example, historical crime data serves as one of the primary components of information that is fed into these surveillance tools.

This is significant because historical crime data is embedded with bias and discrimination which leads to the reinforcement and reproduction of criminal justice and legal biases but on a much wider scale given this expansive and proliferating surveillance architecture. As one report from the Electronic Frontier Foundation finds “Police are already policing minority neighborhoods and arresting people for things that may have gone unnoticed or unreported in less heavily patrolled neighborhoods. When this already skewed data is entered into a predictive algorithm, it will deploy more officers to the communities that are already overpoliced.” Another related issue that would feed into this reproduction of inequality is the unequal distribution and deployment of the physical surveillance technology. They will proliferate in areas already subject to higher police activity (areas that include residents primarily of color and low-income). Allowing for an even wider dragnet over a historically targeted population.

One example of this is the tool ShotSpotter, which is a sensor for gunshots that sends immediate alerts to nearby police units, and its proliferation in certain Chicago neighborhoods. As discussed in this article from The Intercept documenting ShotSpotter’s use in what led up to the recent death of 13 year old Adam Toledo, the author states that “ShotSpotter is operative only in low-income Black and Hispanic neighborhoods and is coupled with software, also sold by ShotSpotter, that guides deployment decisions. The inevitable rejoinder will be: That’s where the crime is. Here, we encounter the circular logic of predictive policing by which supposedly scientific methods yield racist results, as overpolicing of communities of color drives an “evidence-based” dynamic that produces more overpolicing and attendant harms.”

Lastly, these surveillance technologies present a daunting future for civil liberties and rights such as right to privacy, speech, due process, etc. In regard to privacy, given a near total absence of guidelines and regulation, essentially all types of digital data, no matter how identifiable or private, is fair game to be collected, aggregated, and analyzed for any sort of prosecution, raising critical questions about what privacy is protected. The British media scholar John Fiske in the article entitled “Surveilling the City: Whiteness, the Black Man and Democratic Totalitarianism” asserts that surveillance and its affront on privacy is a crucial component of the larger power struggle between the rulers and the ruled in which, “Privacy maintains the area where the less powerful can exert control over the immediate conditions of their lives and bodies, reducing it decreases the localizing power of the weak and increases the imperializing power of the strong, [the ruling class, the state, the totalitarian].” In regard to speech, there are also negative implications for the ability to peacefully dissent against the rule of power structures if there is a wide-scale surveillance architecture monitoring these challenges. This is especially relevant for historically oppressed and marginalized groups who as mentioned already, have been systematically targeted and repressed by law enforcement when attempting to challenge power and attain greater rights.

The increasing reliance on AI-generated algorithms to replace human-led oversight of potentially life-altering events and interactions with police spells grave dangers for society’s most vulnerable as has been documented above. The complete degradation of our most basic democratic ideals and values and the erosion of government transparency and accountability are plausible consequences with the widespread adoption of these kinds of surveillance technologies. Ultimately, this technology is nothing more than a tool. Tools are imbued with the intention of those who create and wield them. They can be designed and/or used for ostensibly beneficial purposes, conversely they can be used for malevolent purposes as well. This is why it is then critical to understand the political and historical context in which the tools are birthed into existence. Given the extensive and current sordid utilization of surveillance tools by its wielder, the police, as a means to oppress and control masses of people, we should analyze every subsequent tool designed for our ‘security’ and ‘safety’ with scrutiny and a critical eye.

Over the past year there has been a heightened level of scrutiny and due criticism, resulting in numerous victories in regard to regulating this police surveillance. Various local and municipal governments around the US have banned predictive policing and facial recognition technologies after experiencing strong pushback from well-organized community coalitions. But this is only a start and we must continue to resist the implementation of this algorithmic oppression in its shaping of group behavior towards dehumanized obedience and conformity. We have to seize upon this energy to stay mobilized, organized, and to articulate our demands for more and more police reform. There is an urgent need to address the worst police abuses and it is incumbent upon all of us to stand in solidarity with those who experience the brunt of this abuse.

This article was posted on Wednesday, April 28th, 2021 at 7:23pm and is filed under Artificial Intelligence (AI), Police, Privacy, Racism.