White privilege is also embedded in the objects of your daily life

The murder of George Floyd in May 2020 sparked a now-familiar sequence of events: massive protests and public statements affirming a truth that is, despite extensive documentation, often overlooked: racism is not merely a series of individual lapses, but a structural phenomenon.

We can see this, and it has been widely denounced, in certain institutions. This is much less the case, however, when we move away from visible institutions (police, justice, employment) to focus on fields perceived as more neutral, particularly science and technology. These fields continue to enjoy a special status: they are seen as objective, neutral, and external to social relations. Here, another idea persists: that technical objects are, by nature, beyond suspicion.

It is precisely this idea that must be questioned.

Scientific Neutrality: A Convenient Fiction

Science enjoys a special status. It is associated with objectivity, rigor, and a form of truth that transcends individual interests and biases. This perception is not merely a matter of trust in scientific methods; it also has a political effect: it tends to place scientific and technical outputs beyond the reach of ordinary criticism. It also allows us to avoid asking certain questions: if technical objects are neutral, then the inequalities they produce can only stem from their uses.

In this context, inequalities are most often thought of in terms of access: who becomes a scientist, who does not, who funds research, who benefits from it. These questions are essential, but they overlook a more subtle and troubling aspect: the way in which technical objects themselves can embody and reproduce social hierarchies.

In other words, it is not just a matter of knowing who does science, but what science and technology do.

When Bias Becomes Infrastructure

Research on facial recognition systems offers a particularly illuminating example. In 2019, a study conducted by the National Institute of Standards and Technology showed that false positive rates were significantly higher for African American and Native American populations than for white populations. The gap is not marginal: it can range from one to ten, or even higher depending on the systems tested.

This is not a technical detail. When these tools are integrated into police or judicial systems, they result in mistaken identifications, unjustified stops, and the arrest of innocent people. And depending on your skin color, the consequences are not abstract.

In response to these findings, one explanation frequently arises: these systems are still imperfect and will improve over time. This interpretation has the advantage of preserving the idea of a fundamentally neutral technology that is merely temporarily flawed. All that is needed, it suggests, is to improve the models, expand the data, and invest further. But this interpretation misses the point.

The biases observed are not merely the result of technical errors. These biases have been known, documented, and criticized for years. Their persistence speaks to something else: they are not corrected because they are not treated as a priority. What is at stake is not just a technical limitation, but an implicit hierarchy of what is deemed worthy of being resolved and what is not.

What these systems reveal is not simply a difficulty in recognizing certain faces. It is a hierarchy, embedded in design practices, of audiences deemed important and those deemed less so.

What the History of Technology Teaches Us

This phenomenon is not new. Long before artificial intelligence, other technologies produced similar effects. For decades, early photographic film—widely used in the mid-20th century—was calibrated to accurately reproduce light skin tones. Black faces appeared underexposed, details were lost, and contrast was poor.

The problem was well known. Yet it was corrected only belatedly, and for reasons that speak volumes about the underlying logic at work. It was not primarily criticism from the affected communities that led to adjustments, but complaints in the 1970s from manufacturers—particularly in the chocolate and furniture industries—whose products were poorly rendered in advertising photos.

This is not an interpretation. It is a fact. Kodak’s former director of research acknowledged it himself: “Black skin was never considered a serious problem at the time.” In other words: the company only resolved the issue when it affected economic interests deemed significant.

But another question sheds even more light on this issue. Why didn’t the communities directly affected bring this issue to the forefront sooner, even though the 1970s were marked by major civil rights struggles?

The answer, documented by Lorna Roth, is brutally simple: “The general public believed that these things were based on science and therefore could not be changed.” [1]

Thus, scientific neutrality does not merely serve to describe the world: it creates a barrier. It makes certain injustices harder to challenge, because they appear technical, objective, almost natural. It is not merely that certain problems go unaddressed. It is that, in some cases, they become unthinkable as problems for those who are their victims.

A consequence of these two trends: what is considered a technical flaw depends less on its scale than on the importance accorded to the people or objects it affects. Thus, technical trajectories are not simply guided by objective constraints, but by value systems.

Shifting Perspectives

Taken together, these examples reveal a pattern. Technologies do not merely reflect the social world; they extend certain of its underlying logics. Through their modes of operation, they reflect priorities, blind spots, and hierarchies. They are not external to power relations: they are one of their expressions.

This does not mean that they are intentionally designed to discriminate. The question of intent, while it may have moral relevance, is insufficient for understanding the observed phenomena. What is at stake involves broader configurations in which technical, economic, and organizational choices converge.

If we accept this, then part of the debate must be reframed. It is no longer merely a matter of identifying problematic uses or malicious actors, but of examining the conditions under which certain technological trajectories become possible, legitimate, and sustainable.

To continue viewing science and technology as neutral spaces is to deprive ourselves of part of the analysis. This leads us to seek the causes of inequalities where they are most visible, while leaving intact the structures that contribute to producing them elsewhere.

Recognizing that everyday objects can embody forms of privilege does not mean attributing intent to them. It means taking seriously the fact that they are the product of social contexts, and that, as such, they can reproduce those contexts. Otherwise, the risk is clear: forgetting that the fight for justice also involves scientific and technical objects, from artificial intelligence algorithms to the molecules we bring to market.

[1] Lorna Roth, scholar at Concordia University: Roth, Lorna. April, 2009. “Looking at Shirley, the Ultimate Norm: Colour Balance, Image Technologies, and Cognitive Equity,” in Canadian Journal of Communication. Vol 34, No. 1, 2009: 111 – 136.