The impartiality that governing artificial intelligence (AI) requires is not a virtue any institution holds. It is what is left once an institution gives up its own stake in the outcome—a financial or political interest in who wins the disputes it is asked to judge.
Magnifica Humanitas, Pope Leo XIV’s recent encyclical on AI, is a broad and pastoral document that discusses integral human development, solidarity, the formation of conscience, and a theology of human dignity in a technological age. In the press, it was read mostly as a warning about the technology: the BBC led with the Pope’s call to “disarm” AI and his image of “new digital slaveries.” However, beyond its warning, Pope Leo makes a structural rather than moral move—he unsettles Catholic Social Teaching’s (CST) own habits. This essay follows that one thread: in the digital sphere, the encyclical locates the apex of power not in the state but in private hands. Subsidiarity—the principle that power belongs at the lowest level able to hold it—exists to discipline such a concentration, not to crown it.
For a century, CST placed the state at the top of the order of competence, restrained from below by families and associations. Pope Leo XIV says the height of power now lies elsewhere: “the highest level is not the State, but rather major economic and technological actors that exercise de facto power over the conditions of everyday life” (no. 71).
This sentence dismantles the reflex the tradition reaches for when power concentrates. Since Rerum Novarum, the answer to concentrated capital has been subsidiarity plus a guarantor—politics, oriented to the common good. That the market cannot regulate itself is not new—Francis had already written that “we can no longer trust in the unseen forces and the invisible hand of the market” (Evangelii Gaudium, no. 204). Leo XIV carries the judgment into the age of AI—”it is no longer possible to rely solely on the ‘invisible hand’ of the market” (no. 163)—and restates the guarantor plainly: “Politics has the task of orientating economies and technologies to the common good” (no. 163).
Both claims stand in the text at once. The encyclical does not reconcile them: it keeps the state as guarantor and names a power above it. This essay draws the tension out; it does not resolve it on the encyclical’s behalf. The two roles can be held apart. The state remains the indispensable guarantor—it makes the rules and enforces them, and no engineered body replaces it there. But the same state is at once the rule-maker for AI, a beneficiary of its revenues, a customer, and the payer of its social costs. The point is not that it is impure; every worldly institution is. It is that one actor holds these four roles at once, so the state cannot also be the impartial judge of disputes that its own stakes shape. The office the state cannot fill is adjudication, not government.
So, the encyclical’s own logic points past the state to a question it leaves open: if neither the market nor the state can supply disinterest, where does disinterest come from?
Disinterest as divestment
The encyclical answers through its own history, almost without meaning to. Commentators have already read it as Leo XIV’s “Rerum Novarum moment”—a social teaching for the AI age. The continuity is real, but it runs deeper than the theme. Leo XIV signed the encyclical on the 135th anniversary of Rerum Novarum, and the lineage matters more than the homage: what links the two texts is not a shared subject but a shared condition of their making.
Rerum Novarum appeared in 1891. The Church had lost the Papal States—its territorial and fiscal sovereignty—in 1870. The first great statement of Catholic Social Teaching, the text that placed persons above “capital and profit,” came twenty-one years after the Church was stripped of the worldly power it had defended for centuries.
The disinterested voice followed the loss of the conflicting interest. This is not a coincidence to apologize for; it is the mechanism. Disinterest was not a virtue the Church possessed at the height of its temporal power. It was the residue of having that power taken away.
Political theology knows the intuition underneath. The post-Constantinian argument, sharpest in John Howard Yoder and Stanley Hauerwas, holds that the Church speaks most truthfully when it does not hold the sword, and that prophetic authority is corrupted by the possession of power. Yoder treated the undoing of the Constantinian fusion of church and empire as a recovery rather than a loss. Hauerwas pressed the same point, that the Church’s first political task is to be the Church, not to run the state.
This is not a lens borrowed from outside and pressed onto the text. Magnifica Humanitas states the same renunciation in its own words: understanding truth “as a gift to be shared, not a possession to be monopolized,” Leo XIV writes, “frees the Church from the temptation of seeking forms of presence based on power” (no. 25). The encyclical’s broader frame is otherwise Thomist and personalist—centered on integral human development, conscience, and grace—and the neo-Anabaptist reading of Yoder and Hauerwas would not fit the encyclical as a whole. But this one thread, power renounced as the condition of truthful speech, is the encyclical’s own.
Neither Yoder nor Hauerwas imagined the argument would find a secular use. Magnifica Humanitas generalizes it. The capacity to judge concentrated power impartially is a function of distance from the stake—and that distance can be lost into existence, as it was for the Church, or established by design.
The confession that earns standing
The encyclical does something rarer still. It audits its own capture. Leo XIV confesses that the Apostolic See once “intervened several times in order to regulate and legitimize forms of subjugation,” that condemning slavery took “eighteen centuries,” and for that delay he “sincerely ask[s] for pardon” (no. 176).
In the register of this argument, the confession is not a pious aside. It is the Church divesting its own interested memory in public, refusing to launder its past as though it had always stood outside the structures it now criticizes. An institution that names the times it was captured has more standing to discuss capture than one that claims to stand above the problem.
The encyclical then hands political theology a secular template. Its list of bodies that “limit the abuse of power and defend the vulnerable”—the Red Cross, “whose operational neutrality ensures compassionate care for all,” the abolition of slavery, the United Nations, the 1951 Refugee Convention—is a catalog of institutions whose authority rests on having no stake in the disputes they judge (no. 123).
Disinterest of this kind is not the absence of commitment. The Red Cross is fiercely committed to the wounded and, for that very reason, takes no side between the armies. That is what the encyclical’s phrase “operational neutrality” names: partisan for the vulnerable, neutral as to who wins. The same was true of the Church after 1870. It did not become indifferent; it lost a competing stake, the temporal interest that had compromised its advocacy, and kept the commitment underneath. Disinterest here means having no stake in the outcome of the dispute—not detachment from the good at stake.
A skeptical reader will object that such bodies are hardly neutral. The Red Cross, the United Nations, and the refugee regime can be read as instruments of a particular order, their “neutrality” a way of making contestable arrangements look natural. The objection is real, but it hits the wrong target. It defeats neutrality as the absence of commitments; it leaves untouched neutrality as the absence of a stake in who profits. These institutions are not neutral about human dignity—they are committed to it and built so that no party to the dispute pays their bills. That is the renunciation worth engineering: not the emptying of conviction, but the removal of the competing interest that would bend it.
The Red Cross, unlike the Church, did not lose its stake by defeat. It was designed without one. If disinterest can be acquired by divestment, it can also be engineered from the start—authority grounded in giving up the competing stake rather than in holding it.
Not integralism
It matters to say what this argument does not license, because the temptation here is integralist and the encyclical refuses it. Nothing in it recommends that the Church govern AI, or that ecclesial authority stand in for political authority.
Leo XIV is explicit: the Church “does not claim to assume the functions belonging to the State” (no. 21), and “does not claim to possess a monopoly on truth” (no. 25). Read this way, Magnifica Humanitas is an anti-integralist document. Its logic does not return power to the Church.
It describes an institution that neither the conflicted state nor the renouncing Church can itself be. The Church models disinterest; it does not take the referee’s chair. That distinction is the whole argument, and to miss it is to turn an institutional insight into a theocratic one—the misreading that the text forecloses.
Building the referee
One objection deserves an answer before the task itself. The encyclical’s center of gravity is not institutional but moral: integral human development, solidarity, a culture of encounter, the formation of conscience. Why, then, isolate the institutional thread from so pastoral a document? Because the moral program does not close the institutional question, it sharpens it. Forming better actors does not remove the conflict of interest: market and state each hold a stake in AI, and even well-formed actors are not disinterested judges of a power they profit from. The question of who can judge survives the encyclical’s deeper answer—that is why this thread is worth pulling out of the shadow of the rest.
Which leaves the task, and it belongs to political theology as much as to policy. The referee that AI requires cannot be the market, which is a player; nor the state, conflicted on four counts at once; nor the Church, which by its own account renounces the office.
It has to be built: a body with no fiscal stake in the technology, insulated from the electoral clock, and durable across the decades over which AI reshapes work, truth, and security. What such a body might look like is genuinely open—an adjudicator built on the Red Cross’s principle rather than the regulator’s, funded so that it gains nothing from the technology’s spread and loses nothing from its restraint. The encyclical does not design it. It establishes that the design problem is the right one.
Designed neutrality is not self-sustaining, and the encyclical’s examples show the strain. The Red Cross, the United Nations, and the refugee regime are exposed to funding pressure, political capture, and the slow erosion of independence; the encyclical’s own word for their achievement is “fragility” (no. 123). So, the harder half of the design problem is not how disinterest is acquired but how it is kept. A body has to be funded, staffed, and renewed so that it neither profits from the technology’s spread nor falls to those it is meant to judge. Distance from the stake must be maintained, not merely established—and what sustains it over decades is partly the culture of conscience and solidarity the encyclical places at its center. Here, the moral program and the institutional one are not rivals but conditions for each other.
What it cannot supply is the account of legitimacy that such a body would require: by what authority an institution, neither elected nor sovereign, may judge the common good. Naming those conditions—neither democratic in the ordinary sense nor theocratic—is work political theology is equipped to do. Magnifica Humanitas has set the question down; it is for the discipline to take it up.
A note on AI use: The author used Claude (Anthropic) to improve the clarity of the English as a non-native speaker and as a sounding board in structuring the argument. The sources, interpretation, and argument are the author’s own; he reviewed and edited all content and takes full responsibility for it.
