Medicine & Health1 February 2026
Generative AI in healthcare: The High-Stakes Control Room
Source PublicationSociology of Health & Illness
Primary AuthorsGross, Geiger
Imagine a massive hydroelectric dam situated above a bustling valley. On the surface, the mechanism seems simple. Water flows in, hits a turbine, and electricity powers the town below. Clean. Efficient. Automatic.
But look closer. If the sluice gates open too wide, the pressure might crack the containment walls. If the water is diverted upstream by a private landowner, the reservoir runs dry. The turbine does not spin in a void; it relies on the geology, the weather, and the engineers in the control room fighting over the levers.
This is the precise mechanism at play with **generative AI in healthcare**. We often stare at the shiny new "turbine"—the algorithm itself—and forget the immense pressure of the water behind it or the politics of the control room. A recent study conceptualises these tools not as standalone gadgets, but as part of a "sociotechnical assemblage". That is a fancy way of saying the machine is glued to the people who built it, the laws that govern it, and the money that fuels it.
The researchers, working alongside social justice organisations, suggest that this technology is currently being forged in the high-friction zone between three spinning gears:
1. The Regulatory Sphere: The inspectors trying to ensure the dam does not burst.
2. The Market Sphere: The investors who want to sell the electricity for maximum profit.
3. The Healthcare Delivery Sphere: The townspeople who actually need the lights on to perform surgery.
If the Market gear spins too fast, it strips the teeth off the Regulatory gear. The study observes that current development is being "motored" by two opposing forces. On one side, you have data capitalism—the drive to extract value from patient data like mining coal. On the other, there is data justice—the push to ensure the benefits do not bypass the people who provided the data in the first place.
If we let the market dictate the flow, the study suggests we risk a system designed solely for enrichment. The algorithms might become highly efficient at billing but terrible at equitable care. Conversely, if civic society grabs the steering wheel, we might direct that power towards public value.
The authors argue we must look behind the "promissories"—the marketing hype—to see who is actually holding the blueprints. The goal is to ensure the dam powers the whole valley, not just the private estate on the hill.
But look closer. If the sluice gates open too wide, the pressure might crack the containment walls. If the water is diverted upstream by a private landowner, the reservoir runs dry. The turbine does not spin in a void; it relies on the geology, the weather, and the engineers in the control room fighting over the levers.
This is the precise mechanism at play with **generative AI in healthcare**. We often stare at the shiny new "turbine"—the algorithm itself—and forget the immense pressure of the water behind it or the politics of the control room. A recent study conceptualises these tools not as standalone gadgets, but as part of a "sociotechnical assemblage". That is a fancy way of saying the machine is glued to the people who built it, the laws that govern it, and the money that fuels it.
The three spheres of Generative AI in healthcare
The researchers, working alongside social justice organisations, suggest that this technology is currently being forged in the high-friction zone between three spinning gears:
1. The Regulatory Sphere: The inspectors trying to ensure the dam does not burst.
2. The Market Sphere: The investors who want to sell the electricity for maximum profit.
3. The Healthcare Delivery Sphere: The townspeople who actually need the lights on to perform surgery.
If the Market gear spins too fast, it strips the teeth off the Regulatory gear. The study observes that current development is being "motored" by two opposing forces. On one side, you have data capitalism—the drive to extract value from patient data like mining coal. On the other, there is data justice—the push to ensure the benefits do not bypass the people who provided the data in the first place.
If we let the market dictate the flow, the study suggests we risk a system designed solely for enrichment. The algorithms might become highly efficient at billing but terrible at equitable care. Conversely, if civic society grabs the steering wheel, we might direct that power towards public value.
The authors argue we must look behind the "promissories"—the marketing hype—to see who is actually holding the blueprints. The goal is to ensure the dam powers the whole valley, not just the private estate on the hill.
Cite this Article (Harvard Style)
Gross, Geiger (2026). 'At the Crossroads of Data Justice and Data Capitalism: How Generative AI in Healthcare Mobilises Its Assemblages.'. Sociology of Health & Illness. Available at: https://doi.org/10.1111/1467-9566.70150