1. A Pope Enters the Machine Age

A question that does not fit on the dashboard

Imagine a public agency preparing to replace a slow application process with an artificial intelligence system. Its staff are exhausted. Applicants wait months. Some give up before receiving help to which they are entitled. A demonstration suggests that the new system can process applications in minutes, detect missing information, and explain the next step in several languages. The people proposing it have something valuable to offer.

Then someone asks what will happen to the applicant whose circumstances do not fit the system’s categories. The presenter explains that the model is highly accurate. Someone asks who will hear an appeal. The presenter shows another chart. Someone asks whether the money saved will improve the service or simply reduce the budget. The room becomes less comfortable.

These questions do not refute the technology. They reveal that its evaluation is unfinished. Speed matters because waiting can hurt people. Accuracy matters because mistakes can hurt people. But the distribution of authority matters because people can also be harmed by a system that performs its assigned task very well. An agency might become extraordinarily effective at excluding those it has decided not to serve.

This is the space into which Pope Leo XIV has stepped. He is not needed to explain how a model performs matrix operations. His presence becomes relevant when the conversation moves from what a system accomplishes to the kind of society its success would create. A faster institution is not automatically a more decent institution. A richer society is not automatically a society in which more people can live securely. An impressive demonstration does not settle either question.

On 10 May 2025, addressing the College of Cardinals, Leo explained that his choice of name chiefly connected him to Leo XIII, whose Rerum Novarum addressed the social question during industrial transformation. He linked the contemporary significance of that inheritance to artificial intelligence and its implications for human dignity, justice, and work.¹ The connection was therefore explicit from the beginning of his pontificate. The machinery had changed. The question of who benefits, who bears the costs, and who can answer back remained.

There is a temptation to make this encounter a drama between two rival costumes: the ancient robe and the modern laboratory coat. That picture is too simple to be useful. A software engineer can be deeply concerned about justice. A religious leader can misunderstand a technical system. A secular government can defend vulnerable people. A church can fail them. No occupation comes with an automatic exemption from error, and no tradition has a monopoly on moral attention.

The more interesting encounter is between different kinds of question. Technical expertise investigates what can be built and how reliably it works. Political judgment asks which decisions should bind a community. Moral reflection asks what we owe one another even when neglect would be profitable. Religious reflection can add a claim about humanity’s origin and ultimate meaning. These inquiries overlap, but none simply absorbs the others.

Leo’s intervention deserves attention because it interrupts an easy transfer of prestige. The people who can build an extraordinary system may have earned admiration for that achievement. They have not, through that achievement alone, earned unlimited authority over the people affected by it. Recognizing this does not diminish engineering. It protects the difference between a contribution to society and a mandate to govern society.

The unfinished argument of the factory

Rerum Novarum, published in 1891, defended private property while rejecting the idea that the labor relationship could be judged solely by what stronger and weaker parties happened to accept. It discussed workers’ conditions, wages, association, and the obligations of public authority. Its approach opposed socialist abolition of private ownership while also challenging exploitation within industrial capitalism.² Those commitments make it harder to recruit Leo XIII as a convenient mascot for either unrestricted markets or an all-directing state.

The document belongs to its time. Its assumptions about family and social hierarchy cannot simply be carried forward as a ready-made answer to every contemporary question. Nor did a papal text by itself settle the conflicts of industrial society. The useful inheritance is an insistence that economic arrangements are arrangements among human beings. They can therefore be assessed by standards beyond the quantity of output they generate.

Consider a hypothetical factory that doubles production after buying new equipment. That improvement leaves several possibilities open. Workers might earn more. Goods might become cheaper. Owners might receive higher returns. Working hours might fall. Competitors might close. The factory might expand and employ additional people, or dismiss half its staff. Nothing in the physical capacity of the machine specifies the final distribution.

The same distinction matters for AI. If a team can complete in one day what previously took five, a new possibility has appeared. What follows depends on decisions about prices, wages, schedules, staffing, ownership, and access. Calling the result inevitable hides those decisions. Calling every gain exploitation hides the real value that increased capability can create. Both errors make it harder to choose well.

A useful technological transition can involve genuine losses. An occupation can become less necessary even when the new service is better. The affected worker does not need to deny the improvement to have a reasonable claim on the society that benefits from it. Conversely, a society should not be required to preserve every inefficient task forever in order to respect the people who once performed it. Respect concerns persons, not the permanent survival of every job description.

This is where the factory analogy illuminates without determining the answer. It directs attention to bargaining power, dependence, and the allocation of gains. It does not prove that AI will reproduce the same employment trajectory as earlier machinery. A tool that changes the cost of cognitive work may affect occupations, institutions, and identities in ways that require fresh analysis. Historical comparison should sharpen observation rather than replace it.

There is another difference. A machine on a production line changes what an organization can manufacture. An AI system may also change what the organization can notice, infer, recommend, or persuade someone to do. Its effects can pass through a person’s interpretation of the world. A system might become the place where someone asks what an official letter means, whether a relationship is healthy, or which claim deserves belief. The stakes include the conditions under which people form judgments.

Yet even here, continuity matters. Powerful institutions have always tried to organize attention and classify populations. The novelty is not that influence has suddenly entered human life. It is that new tools can alter its scale, cost, speed, and personalization. If we exaggerate novelty, we overlook existing protections worth strengthening. If we deny novelty, we may apply old protections to new arrangements without noticing where they fail.

A serious reading of the two Leos therefore begins with a double discipline: preserve the moral question, investigate the changed circumstances. The first prevents technological excitement from dissolving concern for people. The second prevents inherited language from becoming a substitute for understanding what is actually happening.

Disarming without abandoning the work

Leo’s encyclical Magnifica Humanitas is dated 15 May 2026. Its organizing concern is safeguarding the human person in the age of AI. It treats technology as a human activity with beneficial possibilities and dangerous uses, discusses concentrated power, and calls for shared responsibility. It does not present technical development as inherently hostile to humanity.³ That starting point is essential to reading its stronger language fairly.

At the document’s public presentation on 25 May, Leo called for AI to be disarmed. In the same address, he described this as freeing it from purposes of domination, exclusion, and death, and urged people to keep building. He also acknowledged that the Church does not possess the technical answers or seek to replace technical expertise.⁴ Extracting only the demand to disarm would turn a qualified intervention into a misleading slogan.

Still, a powerful metaphor needs careful handling. Nuclear weapons and AI are not interchangeable kinds of thing. AI describes a wide field of methods and applications. The same broad family of techniques may help translate a medical instruction, identify a crop problem, write software, or support a harmful operation. The moral evaluation must reach the purpose, the deployment, and the institutions involved. A metaphor that summons attention should not be allowed to erase these distinctions.

The constructive reading is demanding. It asks us to reduce forms of domination while increasing human ability. That is harder than simply choosing enthusiasm or refusal. A system that helps a patient understand treatment can increase the patient’s agency. A system that quietly prevents the same patient from questioning an automated decision can reduce it. The relevant distinction runs through the uses of the technology rather than between people who like technology and people who do not.

To disarm a harmful relationship might mean removing a secret incentive to manipulate users. It might mean allowing workers to challenge an assessment. It might mean restricting a system’s power to take irreversible actions. It might also mean making a beneficial tool available more widely so that a small number of institutions cannot monopolize its advantages. These are proposals developed in this book, not claims that Leo has endorsed every mechanism or that a single mechanism will work everywhere.

The metaphor also turns back toward the person employing it. A moral institution can speak in ways that invite thought or close it down. It can encourage responsible inquiry or make people afraid to ask questions. If disarming means reducing unjust domination, then religious authority cannot place itself outside the inquiry. It should be prepared to explain how its own interventions protect conscience, permit disagreement, and remain open to correction.

The same applies to governments. Public ownership of a system would not automatically disarm it. A government may possess a legitimate public mandate and still use a tool carelessly, discriminatorily, or oppressively. Private ownership does not automatically make a system unjust either. A business can provide an excellent service under rules that protect the public. What matters is the actual arrangement of purposes, powers, obligations, and remedies.

The invitation to build is therefore more than a reassuring sentence after a warning. It prevents moral criticism from becoming satisfied with its own purity. An ethical position that identifies every danger but refuses to help create better options leaves existing suffering in place. People waiting for diagnosis, education, accessible communication, and reliable public services have interests in progress. They, too, belong in the moral account.

A voice at the table

Why should someone who does not believe in God listen to a pope on AI? There is no need to begin by accepting papal authority. One can examine the argument in the same way one examines a philosopher’s argument, a scientific institution’s warning, or a community’s testimony. Its value depends in part on what it helps us see and whether its conclusions withstand scrutiny.

Religious language can identify a human being as loved before achievement, valuable before usefulness, and answerable for the use of power. These ideas can unsettle a culture that rewards visible performance. They can also produce disagreements. A reader may reject the theological explanation while recognizing a practical danger in making a person’s claims depend entirely on productivity. Another reader may accept the theology yet disagree with a proposed regulation. These are different disputes and should be treated separately.

The public contribution of a religious tradition does not require that every person become a believer. Nor should participation in public discussion require religious people to conceal the convictions that motivate them. The harder obligation comes when an argument becomes a rule enforced on everyone. At that point, officials need to explain the rule to citizens who do not share its religious premises and provide a way to contest its operation.

This does not mean that every rule will satisfy every citizen. No workable society achieves complete moral agreement. It means that disagreement cannot be dismissed by pointing to a speaker’s sacred office, impressive wealth, advanced degree, or large audience. Authority must have a scope. A spiritual leader may guide believers within a religious community. A scientist may have special authority concerning evidence in a field. Neither role alone settles every public decision.

There is a corresponding burden on the skeptic. It is not enough to observe that religious institutions have sometimes failed and conclude that their current arguments need no examination. A valid criticism can come from an inconsistent critic. Indeed, a principle may be useful precisely because it exposes the critic’s own shortcomings. The standard should travel in both directions. If an institution demands accountability of technology companies, it should welcome serious questions about its own accountability.

For believers, this reciprocity need not weaken faith. It can express the difference between the sacredness of a person and the fallibility of the structures through which people act. For nonbelievers, it offers a way to cooperate without pretending that metaphysical differences have disappeared. Shared work does not require a single account of ultimate reality. It requires enough agreement about what will be done, what will be protected, and how disputes will be handled.

On 24 September 2026, addressing the Pontifical Academy of Sciences, Leo explicitly recognized possible advances in medicine, food and energy systems, scientific understanding, creativity, and cooperation. He paired those possibilities with concerns about surveillance, manipulation, discrimination, security, and environmental demands.⁵ The speech supplied another reason to resist portraying his position as a choice between faith and discovery. His argument concerned the responsibilities attached to discovery.

The best response is to make that responsibility specific. If a claim about a system’s danger is wrong, correct it. If its benefits are exaggerated, ask for evidence. If a moral appeal overlooks people who would benefit from innovation, bring them into view. Respectful engagement is not ceremonial agreement. It is the willingness to improve the common understanding, including when doing so requires contradiction.

The scale of the person

At UNESCO in Paris on 25 September 2026, Leo questioned the contrast between attributing intelligence to computational systems and treating people as disposable when they cease to be productive. He connected human dignity to vulnerability, birth, illness, and mortality.⁶ These claims reach beyond a technical dispute about whether a machine should be called intelligent. They challenge a standard by which people themselves are valued.

Imagine two travelers using the same hypothetical airport information kiosk. One speaks the language its voice interface recognizes most easily and can hear its answers. The other has a hearing impairment, reads a different language, and is traveling alone. A system designed around the first person may repeatedly fail the second. If the terminal has removed alternative ways of asking for help, a convenient service has become an obstacle at exactly the point where assistance matters most.

Here moral attention changes the design problem. The question becomes how the service can recognize the obstacles it helps create. A translation tool may be useful. Readable signs may matter more. A staff member may be necessary. The appropriate solution could involve AI, ordinary software, physical design, or better staffing. Starting from the person keeps the institution from confusing a fashionable tool with the whole task.

There is no virtue in making public systems slow merely to keep them visibly human. A kindly employee who gives inaccurate directions is still failing the traveler. An automatic notification that prevents someone missing a connection may perform a humane function without feeling compassion. What deserves protection is the person’s opportunity to receive fair treatment and meaningful help. Human contact matters where it makes those goods possible, not simply as a ritual added to every process.

This view also changes how we understand ambition. Ambition need not mean building a system that can replace every human judgment. It can mean building one that helps more people exercise judgment well. It can mean enabling a small clinic to offer a service formerly available only in a rich city. It can mean giving someone time back, removing needless humiliation, or revealing an error that had become routine.

Such achievements may lack the theatrical appeal of a machine that seems to know everything. Their importance is measured in a different way. A parent understands a diagnosis. A student asks a question without embarrassment. An official finds a pattern of unfair refusals and repairs it. Someone whose life had been reduced to an administrative inconvenience becomes visible again.

The pope enters the machine age, then, with a question that remains useful even to those who reject his answer about God: what must remain true of our treatment of people as our capabilities expand? The question does not belong to him alone. It belongs to every person who might one day be evaluated by a system whose rules they did not choose.

We should welcome greater intelligence when it helps us understand and relieve suffering. We should welcome greater power when it widens the possibilities of a good life. But neither intelligence nor power can by itself tell us who is entitled to command. That requires another kind of account, one that the people living under the decision have a standing to demand.

Sources and notes