Loss, Technology, and Agency
An Introduction to the AssistiveLab Project
This text introduces the AssistiveLab project.
It retraces the biographical and technical trajectory from which this site emerged and offers a reflection on the material and human infrastructures that make agency possible.
I am a sociologist and photographer, a researcher at the CNRS, and I live with limb-girdle muscular dystrophy, a severe and progressive genetic neuromuscular disease.
Neuromuscular diseases have a peculiar trait: their progression unfolds within what still feels like an ordinary bodily experience. Losses rarely appear as dramatic ruptures; instead, they take the form of almost imperceptible shifts. It is difficult to notice a gesture that has become slightly slower, an unusual fatigue, or a movement that has lost some of its precision.
When a gesture begins to slip away, one does not immediately change the entire infrastructure of everyday life. Faced with daily activities, one adapts. One compensates. Habits shift slightly. Existing arrangements are improvised upon: a small wedge here, a piece of foam there; a piece of furniture raised slightly, a support improvised. What remains possible is extended by adjusting the environment to a body whose gestures are becoming more limited, whose range of motion is shrinking. Nothing that yet seems to fundamentally disrupt the order of things.
These adjustments are often discreet, almost invisible. The loss is too.
Yet the disease continues to progress, and disability gradually expands its territory, nibbling away at everyday activities until, one day—sometimes abruptly—one realizes that what was still easily doable a few years earlier is no longer possible at all.
And that is when a threshold sometimes appears more clearly.
I was eleven years old. That evening, I managed to get up from my bed, which had been raised on metal rods. I walked unsteadily toward the television I wanted to turn on. I lost my balance, fell, and fractured my ankle.
A few months later, I stopped walking altogether.
Before that, moving around had already become difficult: a motorized tricycle allowed me to travel longer distances, while walking remained possible for short ones. Switching to a wheelchair marked a turning point.
The wheelchair was the first visible, massive, obvious assistive device. It restored a form of mobility that my body could no longer provide on its own. But it also transformed my relationship to space, to distance, to architecture, and to other people.
Technology compensated. It did not erase the loss. It reshaped the world—my world.
This initial experience reveals something more general: our capacities always rely on material infrastructures that only become visible when the body can no longer mobilize them naturally.
This is particularly true of the activities that now occupy a central place in my life: sociology and photography.
These two practices—producing analysis and producing images—both depend deeply on gestures, tools, and material arrangements. Thinking, writing, framing, triggering the shutter: none of these activities is abstract. Each rests on a precise technical ecology that usually remains invisible as long as it works.
When physical capacities diminish, this infrastructure suddenly becomes visible.
Photography offers a clear illustration.
As the son of a photographer, I had always been intrigued by the practice. Although I quickly lost interest in the camera I was given when I was eight or nine years old, the desire to capture the world around me returned much later. Yet taking photographs required being able to handle a camera, raise it to eye level, frame through a viewfinder, and stabilize the gesture—a sequence of movements I was already unable to perform.
The first digital cameras equipped with articulating screens opened a new possibility: framing images at knee height without having to lift the camera to my eyes.
Around this foundational piece of equipment, a whole series of small improvisations and adjustments gradually emerged. A cardboard box placed on my knees allowed me to support the camera and gain a little more mobility with my legs. I had to find a box with precisely the right dimensions to fit between the armrests of my wheelchair and whose height would match that of my joystick. It needed to be rigid enough not to bend under the weight of my arms, yet light enough for my legs, whose musculature was already limited. The drawer box I eventually chose was neither elegant nor optimal. Yet it allowed me to take a very large number of photographs.
I threw myself almost compulsively into this activity, which offered me a new form of expression, a space for exchange and sharing, and perhaps also an opportunity to show what I was capable of. During the four years of my sociology PhD, this intensive photographic practice almost certainly played a decisive role, offering a parallel path of expression that helped me carry through a particularly demanding research project.
But eventually my improvisations became insufficient. Holding the camera body, adjusting the settings, triggering the shutter—these gestures gradually slipped out of reach. Little by little, my photographic practice declined. I occasionally continued producing images by relying on the arms and gestures of my personal assistants. But as the frustration of a gesture too far removed from my intention began to outweigh the pleasure of taking pictures, photography became rare for several years, despite a desire that had never disappeared.
The robotic arm arrived at that moment—not as a marginal improvement, but as a radical turning point.
This device constitutes an infrastructure. Through its refinement and customization, it allowed me to return to photography as a demanding practice, one that matched the idea I had of it.
Adopting it was not simple. Once again, it required improvisation, experimentation, and the invention of ad hoc solutions tailored to my specific needs. Its very high cost, and the insufficient funding available for such equipment, also significantly delayed its implementation. Yet after a series of trials, adjustments, and refinements, I was not only able to return to photography, but to practice it in ways I had never been able to before. I can now raise my camera above my head or lower it close to the ground to achieve framings that were previously impossible for me.
In a kind of ironic twist—almost a paradox of fate—just as I was experiencing this gain, this rare feeling of reclaiming a bit of ground from the disease, another loss occurred in parallel, one at least as significant and perhaps even more disruptive.
While the robotic arm was allowing me to photograph again, my hands were becoming unable to sustain the gesture of writing.
Once again, the loss was gradual. I modified the fan-shaped layout of my desk. I changed keyboards. I chose a laptop that allowed me to position my hands in a way that preserved as much mobility as possible. When moving my fingers across plastic keys became impossible, I turned instead to the miniature screen of my phone. I tried in every possible way to preserve the gesture for as long as I could.
Then one day, that was no longer enough.
Speech recognition already existed. But in my case it worked poorly: an atypical voice, assisted breathing, an irregular airflow. It promised compensation but rarely delivered. It promised empowerment; in practice it mostly produced wasted time and frustration.
Just as I had stopped taking photographs, I gradually stopped writing. Not entirely, of course. A few words—the most necessary ones—continued to exist. But all the others—the ones that truly matter—slowly disappeared.
It took the emergence of artificial intelligence models for a threshold comparable to that of the robotic arm to be crossed. What the robotic arm did for images, Whisper—the foundational speech recognition model developed by OpenAI—did for text. It made it possible to rebuild a practice that had become inaccessible.
But once again, nothing happens immediately. One must test microphones, compare audio interfaces, normalize vocabulary, design prompt architectures, and develop correction procedures.
The two examples discussed here both involve ambitious technological projects that can still be considered “advanced” technologies today. Yet the best technology is not always the most sophisticated one. The right solution is the one that works in a given situation—the one that integrates smoothly into the ecology of everyday life, that can be understood, adjusted, and repaired.
AssistiveLab emerges from this ambivalent experience of technology: both a source of empowerment and a source of frustration. It opens possibilities while revealing its own limits. It requires adapting, repurposing, and sometimes poaching within the universe of standard objects.
This site exists because a writing infrastructure could be rebuilt.
Two years ago, the text you are currently reading could not have been written. Today it can—so long as the volume of my voice, the quality of my breathing, and the performance of transcription models allow it. This possibility remains fragile, conditional, reversible.
AssistiveLab documents these successive reconstructions: the tiny adjustments that prolong a gesture still possible; the hacks and repurposings of standard objects; the more structured technical devices—mechanical, digital, or hybrid—that allow an inaccessible capacity to be rebuilt; and the methods through which these assemblages can be designed, tested, and stabilized over time.
The goal is not to fuel a fascination with innovation. Rather, it is to make visible, to share, but also to analyze and discuss the concrete fabrication of the material conditions that enable action when ordinary certainties give way. Walking, framing, writing: these gestures then appear for what they are—technical, material, and social assemblages, always situated, always fragile.
These infrastructures are never composed solely of objects and devices. They also rely on forms of human assistance, cooperation, and often invisible labor that allow these technical arrangements to function in everyday life. Autonomy does not mean the absence of dependency. Rather, it refers to the intelligent and equipped organization of dependencies—the ways in which relationships, devices, and practices can be articulated so that action becomes possible.
The point is not to celebrate an individual achievement, but to share methods, detours, and adjustments; to make visible the deeply material and relational dimension of autonomy, while reflecting on the transformations that these mediations impose on the activities themselves. By showing how agency depends on both material and human infrastructures, the project also invites reflection on the very conditions of emancipation.