A photograph that was never meant to exist. There is a photograph that was never supposed to survive. A tintype from 1913, smuggled out of a British prison. It shows a woman strapped to a chair. Her jaw forced open with metal clamps. A tube pushed down her throat.

The woman’s name is Alice Thornton.

She had thrown a stone through a window of Parliament because she demanded the right to vote.(Editor’s note: The image used in this article is a AI-generated reconstruction based on historical sources. It does not replace an original document but serves to visually contextualize a historically documented practice.)

The state’s response was violence disguised as medicine.

When Thornton began a hunger strike in prison, she was force-fed. For weeks. Systematically. She vomited blood. The procedure was officially declared a “medical necessity” — an act of care, according to the state’s narrative. But the photograph tells a different story. It shows a body that is not treated as a subject, but as an object. As a problem to be solved. As an anomaly that must be corrected.

What does this have to do with us?

With the year 2026, with artificial intelligence, with algorithmic justice? Everything.

Measuring the normal

The history of modern medicine is also a history of standardization. In order to understand illness, bodies had to be made comparable. Norms were established: normal weight, normal temperature, normal pulse. These values were not neutral outcomes of observation. They were defined. And whoever defines the norm holds power.

Well into the twentieth century, medical knowledge was based primarily on male bodies. Not out of malice, but because of an assumption so deeply ingrained that it became invisible: the male body was treated as the default, the female body as a deviation. Women were excluded from clinical trials — officially to avoid risks to potential pregnancies. Unofficially because hormonal cycles were considered a “confounding variable” that made research more complicated.

The consequences were profound. Drug dosages tested primarily on male subjects. Diagnostic criteria for heart attacks that failed to capture symptoms more common in women. Crash-test dummies built on male body proportions.

As if half of humanity had been assigned a second-class body.

But gender is only one axis along which this pattern repeats. The same logic appears elsewhere. Black patients are systematically taken less seriously when reporting pain — an echo of centuries-old racist myths about supposedly higher pain tolerance. Older people are more quickly deemed “beyond treatment.” People with disabilities are restricted in their medical self-determination. Trans people struggle for access to treatments that others take for granted.

The structure is always the same: there is a norm — and there are those who do not fit into it.

When algorithms make decisions

This is where artificial intelligence enters the picture, along with its central promise: objectivity.

Algorithms, we are told, will overcome human bias. They will detect patterns we miss. They will decide faster, more precisely, more fairly.

But algorithms learn from the past — and that past is unequal.

An AI system trained to detect skin cancer is fed thousands of images. If those images predominantly show light skin tones, the algorithm learns to recognize melanoma far less reliably on dark skin. The system works — but only for some.

An algorithm designed to predict kidney disease uses historical patient data. In that data, Black patients were referred to specialists less often than white patients — not because they were healthier, but because access to care was unequal. The algorithm learns this pattern and reproduces it as a “recommendation.”

A diagnostic tool for heart attacks is based on studies with very few female participants. It overlooks symptoms that occur more frequently in women.

Technology amplifies what we show it: old exclusions, rendered in new form.

The male body as an invisible norm

There is, however, another side to this story — one that is told less often.

Men, too, pay a price for the normalization of bodies — just in different ways.

Men go to the doctor less often. They speak less openly about psychological distress. They ignore warning signs longer. Not because they are inherently “strong,” but because a cultural script tells them that vulnerability is weakness. Depression in men is diagnosed later, in part because its manifestations often do not align with what is considered “typically depressive.” Aggression, withdrawal, substance abuse — these are rarely interpreted as symptoms of mental crisis.

Here, too, the pattern is clear: the norm does not protect. It constrains.

When a man says “I can’t take this anymore,” he is taken less seriously than a woman saying the same thing — because emotional overload is culturally coded as feminine. When a man becomes a victim of domestic violence, he finds few support structures — because violence against men barely exists in the public imagination.

The gender norm was never designed for anyone. It is a corset that tightens around everyone — just at different pressure points.

What if we asked different questions?

Imagine Alice Thornton had not been restrained. Imagine someone had asked her: What do you need? What is your demand? Imagine we did not start from the norm, but from the human being.

AI has the potential to do more than reproduce old patterns. It could help uncover new ones — if we build it differently. If we ask: Whose bodies are invisible? Whose symptoms go unheard? Whose pain is deemed irrelevant?

Personalized systems — a term that often sounds technocratic — could actually mean systems that see the individual. Not as a deviation from the norm, but as a unique constellation.

That would require more diverse datasets. Studies that reflect real variation. Algorithms that make their foundations transparent. Education that names biases instead of ignoring them.

And it would require space for vulnerability, regardless of gender. Recognition that care is not weakness. Space for pain that does not fit established categories.

The body as a question, not an answer

The question is not: how do we make algorithms fairer?

The deeper question is: how do we create structures in which fairness can exist at all?

That begins with a different way of seeing bodies. Not as datasets to be optimized. Not as deviations from a norm. But as questions addressed to a world that was never built for everyone.

Alice Thornton was force-fed because her body was treated as a problem. Today, we might say: she was asking a question. An uncomfortable one. One that challenged the system itself.

Perhaps that is the most radical insight of all: that every body that does not fit the norm is asking a question. Of medicine. Of technology. Of society.

And that we must learn to listen — carefully.