A few days ago, the Vatican released Magnifica Humanitas, the first major encyclical of Pope Leo XIV, centered on the protection of the human person in the age of artificial intelligence. The timing could not be more urgent.
We have just witnessed the US military use AI to locate and capture Nicolás Maduro, the former president of Venezuela. We have watched AI-assisted targeting contribute to the strike on a school in Minab, Iran, killing 170 people, most of them children. School girls. The daughters, the light of someone's life, reduced to coordinates on a targeting system. Human life flattened into statistics. The use of AI in military intervention deserves an outright ban, and the targeting of civilians should never be permissible under any framework. But we know the United States has a long and documented history of killing innocent people and calling it acceptable loss. One million Iraqis died as "collateral damage" in the hunt for weapons of mass destruction that never existed.
Pope Leo's 200-page document addresses all of this in scope: war, modern slavery, wealth inequality, the erosion of democracy, and the systematic devaluing of human capacity. Some have already begun calling the cultural moment a Butlerian Jihad, borrowing from Frank Herbert's Dune universe, where a holy war was waged not against machines themselves but against the humans who used thinking machines to dominate other humans. The name fits uncomfortably well.
The infrastructure powering this technology is worth examining on its own terms. Data centers are among the most aggressively unremarkable buildings ever constructed: featureless concrete blocks, heavily guarded, humming quietly while they consume entire power grids worth of electricity. That power could have kept thousands of households lit and warm. Instead it generates AI video slop, the kind that floods your feed with algorithmically mutated internet memes nobody asked for, rendered at enormous computational cost, watched once, forgotten.
Beyond electricity, these facilities extract a quieter toll. The land surrounding many data centers has become compromised, with groundwater contaminated to the point where nearby residents rely on bottled water not just for drinking but for cooking and bathing. Surrounding wildlife is displaced or destroyed. The damage is localized, unglamorous, and easy to ignore from a distance.
Then there is the hardware market. These facilities are consuming RAM at a scale that has distorted consumer availability. They have placed orders for memory that has not yet been manufactured, reserving future supply before it exists. The promised abundance of the AI age has, in practice, produced rising hardware prices across the board. The person planning to buy a mid-range phone this year will find less storage for the same price they paid two years ago. Small businesses purchasing workstations face the same squeeze. The billionaires who announced that AI would democratize prosperity are the same people cornering the supply chain.
And then there is Palantir. The surveillance architecture built by Palantir in partnership with US government agencies represents one of the most comprehensive systems of population monitoring ever assembled. It does not only affect American citizens. Its reach is global, its accountability is minimal, and its primary function is to serve the strategic interests of those who fund it.
Here is what makes this complicated: the same underlying technology that enables all of the above also produced AlphaFold.
For decades, predicting how a protein folds from its amino acid sequence was considered one of the hardest unsolved problems in biology. Traditional experimental methods were slow, expensive, and achieved accuracy in roughly the low eighties percentage-wise. AlphaFold solved it computationally, hitting above 90 percent accuracy, and then made the results freely available to researchers worldwide. The downstream effects on drug discovery, on understanding genetic disease, on the study of antimicrobial resistance, are still unfolding. No surveillance, no slop, no military application. Just a model doing science faster than humans could.
This is not a defense of the industry. It is an argument for a third path.
Pope Leo himself does not condemn AI as intrinsically evil. He applies Catholic social teaching as a moral framework for how technology should be designed, governed, and directed. The question is not whether the tool exists but who holds it, toward what end, and at whose expense.
Consider what the same technology looks like when it points the other direction. A microcontroller costs a few dollars. A soil moisture sensor costs less. String them together with an agentic AI loop and you have a system that can grow food through winter in a garage, monitor a small aquaponics setup where fish waste fertilizes leafy greens above them, automate a chicken coop door to open at dawn and close at dusk, regulate water and feed on a schedule, and alert you when something needs attention. Power the whole thing from rooftop solar with a battery bank and you have a food-producing unit that costs a few hundred dollars to build and runs largely on its own.
The step that matters is the one after that. Your neighbor sees the tomatoes. They ask how. You show them. They build their own version and modify it. Over a season or two, a small community forms around shared food, shared tools, and shared knowledge. The grocery run becomes optional. The loop becomes social. This is not a fantasy. Versions of it exist already, scattered across Europe, parts of the Americas, and elsewhere. The technology required is available and affordable today.
What stands between this possibility and its widespread adoption is not technical limitation. It is the same force that always resists decentralized self-sufficiency: concentrated power that depends on dependency.
Magnifica Humanitas is a 200-page document from the oldest continuous institution on earth, written in direct response to a civilizational inflection point. You do not have to be Catholic to recognize what it is saying. AI shaped by the logic of domination produces Minab. AI shaped by the logic of community produces a greenhouse in January, a neighborhood that feeds itself, a future that was supposed to be impossible.
The technology is not the enemy. The question is who it serves.