Measuring Risk: How Data Constructs Reality

I. Introduction

Measurement, like mapping, creates power structures. As Denis Wood writes, maps work because they apply social forces to people and bring into being a socialized space.¹ This notion of measurement as governance parallels the idea that categories like risk or creditworthiness shape human realities rather than portray them neutrally. Systems of data, like systems of mapping, actively produce society rather than mirror it. What makes this especially dangerous is that the people being measured within the maps and other data rarely have their opinions and judgment taken into account in how the categories are built or what they are made to mean.

Ruha Benjamin describes how even something as personal as naming a child becomes politically charged when fed into systems that classify people as risky or safe, because algorithms view names as data.² From airport screening to police databases, these technologies interpret names as markers of risk, again revealing how measurements shape society rather than reflect it. The person being classified is never able to firsthand view the weight their name carries inside the system, thus they have no mechanism available to contest it. That asymmetry — where the system is aware of your life as both qualitative and quantifiable data and yet you simultaneously know nothing of the system — is a calculated tool to maintain authority. Through repetition and institutionalization, society collectively actualizes these systems as truth to sustain order and legitimacy.

II. Risk as a Calculative System

Risk can be defined by calculative systems that quantify social functions today. In Wood's account of the rise of the map with the rise of the modern state, he underscores how mapping emerged to manage territory and ultimately political control.³ Modern systems of measurement are similar in their purpose: to manage populations and codify power. Just as early maps emerged from the state's need to catalog land and people, modern data systems quantify risk to maintain order and control. In other words, when the logic is hidden behind a corporate algorithm or a government database, the ability to push back and resist collapses, and the judgment simply stands as true. That is the difference in which a citizen can see a map on a wall and argue with it, whereas a risk score is invisible, proprietary, and presented as settled.

Benjamin's concept of the "New Jim Code" describes how new forms of measurement such as data algorithms perpetuate the same racial hierarchies as older systems of discrimination, but now with the false front of efficiency and neutrality.⁴ What makes the New Jim Code more durable than what came before is precisely that it does not look like discrimination. It looks like math. And because math carries the cultural authority of objectivity, challenging the outcome requires first convincing people that the process itself was never initially neutral — which is a much harder argument to gain support for. When credit algorithms label individuals as high risk, these labels become mechanisms of governance rather than accuracy. Risk, in this sense, is not discovered by data. It is constructed by it.

III. Representation Becomes Control

Representation has acquired regulatory authority through descriptive data becoming systems of control. Wood further observes that once maps are published, they are accepted as descriptions of the way things actually are, even though they cover up the processes and assumptions behind them.⁵ Algorithmic systems replicate this logic by transforming statistical data and correlations into binding judgments on individuals and communities. The binding nature of these judgments is what separates algorithmic systems from older forms of description. A map can be redrawn. A credit score or deportation flag follows a person across institutions, across years, and across contexts it was never originally designed for, accumulating authority the further it travels from its origin.

Benjamin illustrates that once a digital record is created, whether it is a risk score or a gang database entry, it is repurposed as governance.⁶ And unlike a physical record that can be lost or destroyed, digital governance and data codes are permanent, searchable, and shareable across agencies and institutions with almost no friction. A record created for one narrow purpose can be pulled into an entirely different system years later, applied to a situation its creators did not anticipate, and used to make a decision the person it describes has no knowledge of or power over. The transition from representation to control happens invisibly but intentionally, inside systems that claim neutrality.

IV. The Myth of Objectivity

As a society there must be a collective effort to challenge neutrality and expose the human bias behind systems. Wood demonstrates that maps appear objective but are inherently political tools that serve the interests of the state, and he reminds us that neutrality is a social performance.⁷ The problem itself is not just that these tools serve a certain interest. It is that they are designed to look like they serve no interests at all, which is one of the most powerful political moves available. A tool that announces its bias can be resisted; a tool that successfully performs objectivity puts the burden of proof on the person it harms, who must first convince others that the system claiming neutrality was political before they can even begin to argue that its data was interpreted incorrectly.

Similarly, algorithmic risk assessments wear the mask of objectivity while quietly reinforcing historical inequalities. Benjamin notes that racial bias is often doubled or hidden behind the digital design.⁸ The real error to be pursued is not the presence of prejudice but the intentional concealment of it under the guise of fairness. When bias is made to look like logic, it becomes much harder to name and much easier to sustain. This system is difficult to dismantle because it gives institutions plausible deniability while producing the same outcomes. The harm is measurable, but the mechanism is obscured. Moreover, as long as the encoded bias stays hidden, the institution can point to the algorithm and declare the decision was not theirs to make.

V. Feedback Cycles and Self-Fulfilling Inequality

Feedback cycles of measurement enforce control, creating self-fulfilling systems of inequality justified through efficiency. Maps become governing devices that command compliance through their perceived truth.⁹ Humans obey algorithmic scores and conclusions because they appear legitimate, not because they are transparent. There is no need for direct coercion when power lies in belief.

Berger and Luckmann emphasize that routine actions create institutional legitimacy.¹⁰ The process of quantifying risk demonstrates habitualization at scale — repetitive data modeling that feels rational but is historically contingent.¹¹ These patterned choices become institutions sustained by habit and consensus, which then legitimizes data as truth rather than interpretation. Repetition of the data works more effectively than force.

VI. Conclusion

Maps and models are engines that convert human complexity into the language of governance, a process that now demands more attention, accountability, and resistance. Benjamin underscores that the real danger is not the presence of bias but its concealment, and the way systems of data quietly do the work that older systems of discrimination once did openly.¹²

True data ethics demands slowing the pace of innovation to consider who bears its costs. Systems that encode risk must be reimagined to include community input, not simply more sophisticated machinery. Wood's central insight holds: as long as we accept the map as the territory, we will keep mistaking a tool of governance for a fact of nature.¹³