Doe Machines Have a Gender? Notes for Bodies Out of Place
Doe Machines Have a Gender?
Notes for Bodies Out of Place
by Ro
Introduction
A body does not end at the skin. It continues in the name written on official documents, in the drugs that alter its chemistry, in the labour required to keep it alive, in the images through which it learns to desire itself and to be desired. Today, it also continues in profiles, in apps that track its menstrual cycle, in photographs archived online, in late-night searches, in data produced almost without noticing.
This does not make the body any less real. On the contrary, it forces us to ask how many relations its materiality is made of.
Feminist thought has long shown that what we call the “natural body” is never an innocent starting point. Sex, gender, desire, reproduction, labour, and care have been held together by institutions, forms of knowledge, images, and technologies that have made historically constructed relations appear necessary. The digital now enters this history, but does not simply repeat it. It introduces new ways of seeing, classifying, and relating bodies; at the same time, it makes possible new forms of recognition, experimentation, and transformation.
A box that allows someone to select “non-binary” can be the very concrete result of a struggle against the obligation to exist only as either a man or a woman. That same box is also information that a system can record and use. The two do not cancel each other out.
This article starts from this ambivalence. Not to denounce technology as a new patriarchal machine, nor to celebrate every new possibility of representation as liberation. The question is more precise: what happens to bodies when they become at once freer to differentiate themselves and more precisely legible?
Perhaps this is where the question “Do machines have a gender?” can be posed differently. Not only by asking which biases have entered machines, but what bodies we are learning to become together with them.
Gender has always needed to make the body legible.
For centuries, a certain anatomy had to signify much more than itself: a disposition towards labour or care, a sexuality, a reproductive capacity, a position within the family, even a character. The feminised body was associated with care; care was materially assigned primarily to feminised bodies; repetition produced skills, postures, sacrifices, and ways of organising time and life. All of this could eventually return to the body in the form of evidence: women take care of others because they are supposedly, by nature, more inclined to do so.
The norm produces the regularity that it subsequently presents as nature.
A decisive part of feminism has broken precisely this circularity. If the body does not already contain its own destiny, it becomes possible to separate what had been made to coincide: sex and gender, sexuality and reproduction, desire and family, anatomy and social function.
When Donna Haraway proposed the figure of the cyborg in the 1980s, she radicalised this fracture further: no return to an original body, uncontaminated by technology, could solve the problem. The boundary between organism and machine was already political, unstable, traversed. Forty years later, we do not need to imagine science-fiction scenarios to encounter that body.
A body menstruates together with an app. It alters its chemistry through contraceptives, antidepressants, or hormones. It constructs memory through digital photographic archives. It looks for a partner through an interface. It learns a gesture from a tutorial. It sees itself through images that other bodies have produced and that a system has selected.
The body does not simply use technology. It is also formed through the relations it enters with it.
This becomes particularly evident when the body is transformed into data. Face, voice, location, menstrual cycle, sleep, purchases, photographs, relationships, pornography, desires, browsing history. The digital body does not coincide with the profile we intentionally construct: it includes the traces produced as we live.
But a piece of data is not a fragment of reality that enters the computer intact. To record something, you have to decide what matters
An app that tracks the menstrual cycle selects variables, establishes intervals, compares symptoms and recurrences. A dating app makes desire partially translatable into age, gender, distance, preferences, and acts of selection. A form that asks for gender has to establish which answers can exist within it. The machine therefore does not simply encounter difference. It has to give it a form with which it can operate.
This also shifts the problem of inclusion.
If a form only allows a choice between man and woman, adding other possibilities matters. For those who do not exist within that dichotomy, being nameable is not a cosmetic concession. A word can change the relationship one has with one’s own body; a shared category can produce community; an administrative possibility can prevent very concrete violence.
But adding another box does not settle the political problem. Because being included means, at that precise moment, having become legible to the system. The same word can belong to the genealogy of a struggle and to the architecture of a database. This is a more interesting situation than a simple opposition between liberation and control. A category does not lose its value because it can be recorded. Nor does it become emancipatory simply because it expands the possibilities of self-representation.
The question is what happens next.
Computation, in fact, does not only need differences. It looks for relations between differences. Frequencies, similarities, associations, probabilities. And this is where it encounters gender again.
Gender, too, has produced regularities. Not because masculine and feminine behaviours necessarily exist inscribed in bodies, but because a society that distributes labour, money, violence, care, and freedom differently inevitably ends up producing different behaviours.
We can return to care. Care is work, even when it falls outside what is recognised and measured as productive labour. If it is primarily women who reduce their paid working hours to perform unpaid or underpaid care work for children or family members, that division of labour produces observable effects. Careers, incomes, consumption, schedules, and movements bear its traces. When these traces become data, a system can recognise a regularity without knowing anything about the political and economic organisation of labour that produced it.
The data can preserve the effect and lose the conflict.
For this reason, algorithmic sexism does not become interesting only when the machine gets something wrong. Sometimes the machine correctly sees an unequal world. It does not have to claim that women are naturally predisposed to care. It is enough for the past to contain enough women who have performed care work for that relation to acquire predictive value.
The machine loves genealogy because genealogy enables prediction. It is enough for the before to contain enough of the after.
But that “after” has become much less docile.
The digital circulation of words, images, and communities has helped make practicable forms of gender that older sequences could not contain. This is not only a matter of greater visibility for identities that already existed, perfectly formed, waiting to appear online.
Sometimes a word produces something.
Encountering “non-binary,” “genderfluid,” “trans*,” or any other term capable of temporarily adhering to an experience can change the way one reads one’s own past. An image can make a modification of the body desirable. A community can transform a private feeling of inadequacy into an experience that can be shared. A tutorial can transform knowledge into a practicable technique.
The symbolic, here, is perfectly material. It passes through the body and modifies it.
It would therefore be a mistake to read the contemporary proliferation of gender as mere identity inflation or commercial invention. More words can mean more possibilities of existence. The multiplication of signs has expanded the field of what can be thought, desired, and lived.
But it is precisely here that the question becomes more subtle.
Because the proliferation of difference takes place within the same environments that continuously transform practices into information. One searches for a word. Looks at an image. Modifies a profile. Selects whom one wants to meet. Enters a community. Searches for a drug. These are experiences that can have profound consequences in a person’s life; they are also actions capable of leaving traces.
The body can therefore become at once more differentiated and more informative. There is no need to choose which of the two conditions is the authentic one.
It is precisely their simultaneity that characterises part of the contemporary experience of gender.
Informational capitalism, from this perspective, does not necessarily require uniform subjects. It can also function through differentiation, personalisation, variation. Tastes, aesthetics, desires, and identities can be lived experiences and, at the same time, produce information. A category can be the result of a struggle and become a variable. An aesthetic can disrupt a norm and circulate as a commodity. A community can produce belonging within a commercial platform.
None of these transformations automatically cancels the others out.
This is why even the language of “capture” risks being too simple. It assumes that a pure and subversive difference exists first, and that only afterwards does the market arrive to take possession of it. But bodies do not form outside technical, economic, and symbolic relations only to enter them at a later moment.
The cyborg has no paradise before the machine.
And yet it does not follow from this that every hybridisation is liberatory. Perhaps this is where a technofeminist and queer perspective has to become more demanding.
If, at one time, the impossibility of classifying a body could reveal the inadequacy of the norm with particular clarity, today a growing part of difference can be accommodated within much more flexible systems. A platform can add categories. A market can differentiate products. An algorithm can learn new associations.
Mutation does not necessarily have to be prohibited. It can be made operable.
This changes the relation between difference and subversion. It does not eliminate it: it complicates it.
Being legible does not mean being free. But being illegible does not automatically mean being subversive. A politics of gender therefore cannot measure its own radicality solely by the number of identities recognised, nor seek it in the most fluid, incoherent, or impossible-to-name body. Both solutions would continue to place classification at the centre.
The problem emerges instead when we look at what a classification makes it possible to do.
A person can change the gender on their profile: can they change it just as easily on their official documents? A word can make a transition thinkable: what economic, healthcare, and relational conditions make it possible to undertake it? A representation can free a body from shame: does it also change the conditions under which that body works, receives care, desires, and builds relationships?
There is no single answer. And it is important that there is not.
Because the point is not to reconstruct a hierarchy between “real,” material transformations and merely symbolic transformations. A name can materially change a life. An image can produce a desire. A digital community can make possible a relationship that changes the way one inhabits one’s own body.
The material and the symbolic are not two planes that occasionally meet. They continuously produce one another.
Perhaps this is what still makes the body such a powerful point of observation. Not because it preserves a truth prior to technology, but because it is the site where very different relations acquire consistency: a word becomes posture, an image desire, a norm working time, a molecule transformation, a datum prediction.
The body does not stay in its place because it does not have only one.
It is biological and technical, individual and relational, opaque and legible. It passes through the machine and is traversed by it. It produces information that returns to organise the environments in which it will continue to become other.
The question “Do machines have a gender?” can therefore be deliberately left open.
Perhaps the problem is not to discover whether the machine is masculine, feminine, or something else. It is to observe the work through which bodies and machines continue to produce differences, transform them into regularities, and put them back into circulation.
Between the difference that opens up a possibility and the regularity that tries to anticipate it, an unstable space remains.
It is there that gender continues to move.
