We are not our data, but we are treated as if we are
Data are a translation of our complex reality into numerical values. That is a necessarily simplification of the world. It makes the abstract concrete, the intangible tangible. At least, that is the idea. And it is a valuable idea, because data allow us to do calculations and make the complex world intelligible.
What we measured, we know
We cannot rank, prioritize, or assess people in all their complexity. But if you can assign a numerical value to people, that suddenly seems possible.
Sometimes we forget that what we measure is not reality itself. It is a representation of reality. This representation does not come about arbitrarily. We make choices about what we measure, but also how, when, where, how often, and by whom. These are variables that can make a difference to the resulting model of reality.
Therefore it is also an illusion that data are an objective representation of the world. This representation often says something about the world, but just as much about how we see the world. We try to divide the world into categories: male or female, young or old, trustworthy or not trustworthy.
Lets look at a simple example. The length of an object is hard to dispute. But if the measuring instrument is imprecise, you still end up with a distorted picture of reality. An object can also change in size over time. Whether this matters, depends on the context in which you take the measurement. Misrepresenting a person's height by a few millimeters is not a serious problem. However, in aircraft maintenance, precise measurements of materials do matter, because a difference of a few millimeters can have unforeseen consequences.
Length is still something tangible. But lets look at something more abstract, like income. This cannot be directly observed in the world. The word "income" is itself a category, an umbrella term for something. What we mean by income is not a number in itself - that number alone tells us very little. 3000 euros of income means something very different in the Netherlands compared to another country. Income is really an approximation of what a person can afford in the world. In rental situations, we want to know whether someone can pay the rent. In benefits assessments, we want to know whether someone can make ends meet without support. Without context, without interpretation, data are meaningless and without value. It is only within a particular context that they acquire meaning.
Another example: a zip code. In one context, simply an indication of a geographical location. But in another context, it can be interpreted as a predictor of reoffending. The same value, but an entirely different meaning.
"We are not our data, but we are treated as if we are - out of convenience and efficiency."
Centralization
In an era of data-driven working, it may seem logical to make one organization responsible for data management. That organization collects the data, monitors quality, and keeps everything up to date. The advantage seems clear: fewer inconsistencies, fewer trivial errors, one standardized way of organizing things. For citizens, centralization means they only have to submit their details once. For the government, it means decisions can be made more quickly, because all the necessary information is immediately available.
But that advantage is also the disadvantage. Because data come from a single source, the context in which they were generated is lost. Those who use the data also have no clear picture of how the data came to be. Different organizations use data for different purposes. For one organization, income means only a person's monthly salary; for another, it also includes any benefits or allowances they receive. In one context, the fact that someone has received a benefit has no consequences; in another context, that benefit causes the person to be wrongly rejected when applying for support.
When data are detached from the organizations that actually use them, something is lost. Those organizations find it harder to deliver tailored solutions. They also find it harder to explain how they reached decisions, because those decisions are based on data supplied to them according to standards determined elsewhere.
In addition, the original purposes for which the data were collected become blurred. Under the dutch interpretation of the General Data Protection Regulation, data collected for one purpose may not simply be used for another. But when data management is centralized, the same data are used for multiple purposes: making the principle of purpose limitation difficult to uphold.
People whose data do not fit neatly into existing categories, face particular difficulties. Supplying this data becomes harder, because it does not fit tidily into the predefined boxes. A self-employed person with a fluctuating income, someone with a complex family situation, a person whose life simply doesn't fit the available boxes: they fall through the cracks.
"We can now easily find out all sorts of things about people without involving them at all. Just because we can, does not mean we should."
Data separated from the person
Data should not be the leading principle. When you store information about people, you see that data separately from the person themselves. In the past, finding out about someone meant knocking on their door or picking up the phone. There was direct contact. Now, with existing data, all kinds of conclusions can be drawn about someone without that person knowing they are being investigated. They cannot defend themselves or raise an objection until they experience the consequences.
What does it mean that our data are viewed separately from ourselves? We are not our data, but we are treated as if we are - out of convenience and efficiency. This separation between person and data representation has far-reaching consequences for how we approach and treat people. We can now easily find out all sorts of things about people without involving them at all. Just because we can, does not mean we should.
A deliberate choice
The government considers data centralization worthwhile and accepts its drawbacks. It prioritizes efficiency for the majority, even if it becomes harder for those who do not fit the system. But the government exists to serve everyone. And so the government owes it to these people not to leave them behind.
In concrete terms, this means we must design systems that accommodate exceptions - and do not treat them as an afterthought. A human reviewer should always be available for complex cases. There must be transparency about how citizens' data are processed. Organizations must be able to account for how they interpret data. And above all: we must not mistake the map for the territory. Data are a representation of reality. Not reality itself.