Who Owns the Record?
What a papal encyclical on artificial intelligence has to say to records and information managers — and why it matters more than you might think
On May 15th, Pope Leo XIV issued an encyclical — a formal letter of the highest doctrinal weight — addressed not only to Catholics but to “all men and women of goodwill.” Its subject, officially, is artificial intelligence. Its real subject is something records managers have been arguing about for decades: who controls information, who bears responsibility for decisions made with it, and what obligations we owe to the people whose lives those records represent. The document is called Magnifica Humanitas — the Grandeur of Humanity. It turns out to be one of the more rigorous frameworks for thinking about information ethics published this year, in any genre.
The document nobody in RIM is reading (but should be)
Let’s be honest about the usual reaction when a religious text enters a professional conversation: polite acknowledgment, then a determined pivot toward something more immediately applicable. This would be a mistake here. Magnifica Humanitas is a 52-page, 245-paragraph document with 224 footnotes that draws on philosophy, economics, labor law, communication theory, and moral theology to construct an unusually coherent account of what goes wrong when information systems serve power instead of people.
It was issued on the 135th anniversary of Rerum Novarum — Leo XIII’s 1891 encyclical that first applied Catholic social teaching to industrial capitalism, the document that put the Church formally on record about labor rights, fair wages, and the dignity of workers. The choice of date is not incidental. Leo XIV is explicitly framing artificial intelligence as the new industrial revolution — a transformation equally profound, equally capable of liberation or exploitation, and equally in need of a moral framework that goes beyond what markets and regulations alone can provide.
For the records and information management profession, the timing is useful. We are in the middle of an urgent and unresolved argument about AI in RM practice — about automated classification, algorithmic disposition, AI-generated records, and the governance of data that describes real people. Magnifica Humanitas does not resolve those arguments. But it provides a vocabulary and a set of principles that cut through a great deal of the evasion that has attended them.
The encyclical in brief. Magnifica Humanitas (Latin: “The Grandeur of Humanity”) was issued May 15, 2026, by Pope Leo XIV. It addresses the social, ethical, and spiritual dimensions of artificial intelligence, framed through the principles of Catholic Social Doctrine — human dignity, the common good, subsidiarity, solidarity, and social justice. The document explicitly extends these principles into digital environments, including data governance, algorithmic accountability, labor conditions in AI supply chains, and the use of AI in warfare. At 52 pages, it is comparable in scope to major modern encyclicals like Laudato Sì (2015) and Fratelli Tutti (2020).
Babel versus Jerusalem: what kind of system are we building?
The document is organized around two biblical images, and they are worth understanding because they frame every specific argument that follows.
The first is the Tower of Babel — the great construction project of Genesis 11, where a unified humanity pools its resources to build a tower “with its top in the heavens.” The tower fails, the language fractures, the people disperse. Leo XIV reads this not as a story about divine jealousy but about the internal logic of certain kinds of systems: ones built on uniformity rather than communion, on efficiency rather than relationship, on the ambition to dominate rather than the willingness to serve. A project conceived, the encyclical says, with “a uniformity that eliminated diversity” and that “chose homogenization over communion.”
The second image is of Nehemiah rebuilding the walls of Jerusalem after the Babylonian exile. He prays first, then plans carefully, then assigns each family its own section of wall. He listens to concerns, addresses opposition, and coordinates without controlling. The city is rebuilt not through the initiative of one powerful actor but through what the encyclical calls “shared responsibility” — every household, every trade, every age group contributing to a common project.
Applied to information systems, the distinction is clarifying. A Babel system is one where a small number of actors control the infrastructure, set the rules invisibly, optimize for their own metrics, and present the result as neutral and inevitable. A Jerusalem system is one in which governance is transparent, affected communities have a genuine voice, the purpose is explicitly the common good, and accountability is traceable to specific human beings making specific choices.
Most of the AI-assisted RM systems being sold to organizations right now are, in this taxonomy, Babel projects — not because they are malicious, but because structural incentives push toward opacity, uniformity, and the concentration of decision-making authority with the platform provider. The question the encyclical keeps asking is: which city are we building?
“The primary choice is not between a ‘yes’ or ‘no’ to technology, but rather between constructing Babel or rebuilding Jerusalem; between a power that claims to dominate the heavens and a people who work together in the presence of God to rebuild the walls of fraternal coexistence.”
Data as a common good: the argument that changes everything
The most significant single move in the document, for our purposes, is the formal extension of the principle of the universal destination of goods — a longstanding doctrine that the earth’s resources belong ultimately to all humanity, not to whoever happens to own them — to include “patents, algorithms, digital platforms, technological infrastructure and data.”
This is not a minor adjustment. The universal destination of goods has been part of Catholic Social Doctrine since the Middle Ages and has been developed systematically since the 19th century. It does not abolish private property — the encyclical carefully affirms property rights. But it subordinates those rights to a higher obligation: that property be used in ways that serve the common good, not merely the owner’s interests. Applied to data, the argument is that information — including the records that organizations hold about individuals, communities, and public processes — cannot be treated as simply belonging to whoever collected it.
The encyclical makes the policy implication explicit: “ownership of data cannot be left solely in private hands but must be appropriately regulated. Data is the product of many contributors and should not be treated as something to be sold off or entrusted to a select few.” It even calls for thinking “creatively” about managing data “as a common or shared good, in a spirit of participation.”
For records managers, this reframes a question usually posed in regulatory terms — GDPR, FOIA, HIPAA — as a matter of justice. The right of an individual to access, correct, or delete the records an organization holds about them is not a courtesy granted by regulations; it is grounded in human dignity itself. Public records access is not an administrative accommodation; it is a requirement of the common good. Organizational claims to own data generated by or about the people they serve deserve far more scrutiny than they typically receive.
The Tolkien problem: on power, good intentions, and the illusion of the neutral system
J.R.R. Tolkien appears in paragraph 213 of a papal encyclical — surely one of the stranger sentences one can write in 2026 — in a passage about individual responsibility. Leo XIV quotes the wizard Gandalf speaking to the Fellowship: “It is not our part to master all the tides of the world, but to do what is in us for the succor of those years wherein we are set, uprooting the evil in the fields that we know, so that those who live after may have clean earth to till.”
It is a good choice of text. Tolkien was, among many other things, a writer obsessed with the corrupting effects of concentrated power — particularly the subtle corruption that comes not from obviously evil intentions but from the conviction that one’s own good purposes justify acquiring power over others. The One Ring does not tempt the Fellowship with cruelty. It tempts them with the promise of doing more good, more efficiently, at a greater scale.
The encyclical invokes Tolkien in the context of geopolitics and war. But the logic applies with equal force to the information management profession’s current relationship with AI. The appeal of automated records management is not that it serves any obviously bad purpose. It promises to do the work of classification, disposition, and search at a scale and speed that human beings cannot match — freeing us for higher-order work, eliminating the backlog, and solving the problem of unmanaged information growth.
What the encyclical asks — following Tolkien’s logic — is: what is the system optimizing for, and whose values are embedded in that optimization? “Every technical tool embodies choices and priorities through what it measures, ignores, and optimizes,” the document states. “We cannot consider AI to be morally neutral.” The classification algorithm that routes a personnel file to disposal is not neutral. The retention schedule that a machine learning model infers from historical disposal patterns is not neutral. It reflects the judgments of whoever built and trained the model, which in turn reflects the historical biases of the records used to train it.
The “fields that we know” — Tolkien’s phrase — are the systems, policies, and governance frameworks that records managers directly influence. The encyclical’s point, and Tolkien’s, is that this limited scope is not a failure of ambition. It is the actual site of ethical responsibility.
Six things the encyclical actually requires of information governance
Moving from principle to practice: the encyclical’s arguments generate at least six specific challenges for the RIM profession. These are not speculative extrapolations. They follow directly from the document’s positions.
I. The AI decision-provenance record. The encyclical requires that all systems used in consequential decisions “guarantee the possibility of retracing and reconstructing decision-making processes.” For RM, this means any decision assisted by AI — classification, disposition, access control — must generate a documentary record that includes the model version, input data, output, and human review point. This is not standard practice. It should be.
II. Contestable classification. The encyclical requires that decisions affecting people’s “employment, credit, access to public services or reputation” be “understandable, contestable and subject to oversight.” When AI assigns sensitivity levels or retention periods to records about individuals, those individuals — and the records managers accountable for their files — need a meaningful pathway to contest and correct those decisions. Most current automated RM systems provide no such pathway.
III. Provenance tagging for AI-generated content. The encyclical warns that AI “dramatically amplifies” disinformation and that the integrity of the information ecosystem requires verification and transparency. A memorandum drafted entirely by a large language model and signed by an executive is evidently different from one the executive wrote. Current metadata standards do not systematically capture AI involvement in record creation. They need to.
IV. Data rights as pre-legal obligations. The encyclical frames the right of individuals to control records about themselves as grounded in human dignity, not merely in regulation. This means that GDPR compliance, FOIA responsiveness, and right-of-access programs are not optional accommodations that organizations provide as regulatory minimums — they are expressions of a moral obligation that exists regardless of what the law requires in any particular jurisdiction.
V. Community and Indigenous data sovereignty. The encyclical’s concern with the extraction of community-level data — health records, epidemiological profiles, demographic information — as a form of neo-colonialism anticipates the emerging framework of Indigenous data sovereignty. Records about communities belong, in some meaningful sense, to those communities. Governance frameworks for archival holdings that include such records need to reflect that obligation.
VI. AI procurement as a labor-records issue. The encyclical describes in detail the invisible labor of data annotation and content moderation that underlies every AI system. When an organization procures an AI records management tool, it is implicated in the labor conditions associated with it. Responsible procurement requires documentation of training data sourcing and annotation practices — a new category of due diligence records that most organizations are not currently maintaining.
Why the archive is a moral institution, not just a technical one
One of the encyclical’s quieter but most important arguments concerns memory. It notes with alarm “a disconcerting loss of historical memory, as first-hand accounts of the Holocaust and the two World Wars are disappearing. This leads to a selective or distorted rewriting of the past, in a context where fake news and the manipulation of narratives obscure the lessons that have been learned.”
This is an archival argument. The encyclical is saying that the preservation of primary evidence — particularly testimony from victims and survivors of atrocity — is not an act of bureaucratic tidiness but a moral obligation to future generations who will need that evidence to resist the normalization of violence. “Without a living memory of the horrors of war,” the document continues, “political decisions risk being made on the basis of power alone, without any consideration for the long-term consequences.”
The implication for the profession is significant: archives are not passive repositories. They are active participants in the moral life of their communities. Decisions about what to preserve, what to dispose of, what to digitize, what to make accessible, and what to restrict are decisions about what future communities will be able to know about the past. Framed this way, appraisal is not a technical function; it is an ethical one.
The encyclical’s formal apology for the Church’s historical complicity in slavery — naming specific papal documents by title, acknowledging the delay with which the institution came to condemn a practice it had once legitimized — is itself a model of what it means to take archival accountability seriously. You cannot apologize for what you have not documented. You cannot learn from the past if the records of past wrongdoing have been disposed of, made inaccessible, or never created in the first place.
“Without a living memory of the horrors of war, political decisions risk being made on the basis of power alone, without any consideration for the long-term consequences.”
Subsidiarity: Who gets to decide what the records say?
The encyclical applies the principle of subsidiarity — that decisions should be made at the level closest to the people most affected by them — specifically to digital governance. It insists that information governance processes “not be imposed from above in an opaque and unilateral manner” and that communities and intermediary organizations must have “a voice and contribute to the discernment of choices that affect people’s daily lives.”
In enterprise records management, this challenges the standard model of centralized governance. RM programs typically set retention schedules, classification schemes, and access controls at the organizational level and impose them uniformly across business units and communities. This is efficient. It is also, in the encyclical’s framework, a form of paternalism that silences the people most affected by those decisions.
What would a subsidiary model of records governance look like? It would involve genuine participation by the communities whose records are being governed — not just consultation but co-design of retention schedules for records that primarily affect community members. It would mean that algorithm-driven classification systems could be examined and contested by the people they categorize. It would mean that Indigenous communities have meaningful governance rights over archival holdings that document their history. None of this is impossible; some archives are already moving in this direction. The encyclical provides a principled basis for moving much faster.
A resource for the profession
The records and information management profession has a persistent problem that is distinct from the technical challenges it faces. The technical challenges — managing unstructured data at scale, governing AI-assisted systems, preserving born-digital records for the long term — are hard but tractable. The persistent problem is different: it is the difficulty of making the ethical case for records work against short-term cost pressures, against the convenience of deletion, against the institutional temptation to let the archive atrophy.
Regulatory frameworks help — GDPR, FOIA, ISO 15489, and various national archival statutes. But regulations have a ceiling. They establish minimum compliance thresholds; they do not inspire. And they are constantly subject to weakening by the political winds of the moment. A framework grounded in human dignity and the common good is more durable. It makes the argument not in terms of what the law requires but in terms of what we owe each other as human beings living in communities that depend on shared memory.
Magnifica Humanitas does not resolve the open technical questions of AI in records management. It does not tell you which retention schedule to apply to a Teams channel or how to handle the synthetic media problem in an institutional repository. But it provides a vocabulary — accountability, transparency, dignity, subsidiarity, the common good — that the profession already uses, and it grounds that vocabulary in one of the most systematically developed traditions of social ethics in human history.
The Magnificat — Mary’s hymn from which the encyclical takes its title — sings of a God who scatters the proud, puts down the mighty from their seats, and exalts those of low degree. It is a song about power and its reversal — about whose perspective counts, whose voice is heard, whose story gets told. Archives, at their best, do the same work: they give permanence to voices that power would rather erase, and accountability to institutions that would prefer to be forgotten.
That is not a small thing. And it is, the encyclical insists, not a bureaucratic thing. It is a profoundly human one.
Sources and key passages
All citations are to the numbered paragraphs of Magnifica Humanitas (Leo XIV, 15 May 2026), available in full at the Vatican website. Encyclical paragraphs are numbered sequentially and cited by convention as §[number], functioning similarly to verse citations in scripture.
Data as a common good: §§65–67, §108
Algorithmic accountability and contestability: §§102–105, §164
AI decision-provenance and the chain of responsibility: §§199–200
Truth, disinformation, and the integrity of the record: §§132–134, §137
Data colonialism and individual data rights: §178
Supply-chain labor and AI procurement: §§173–179
Historical memory and the archival obligation: §191
Subsidiarity in digital governance: §§71–72
The Tolkien passage: §213, quoting J.R.R. Tolkien, The Return of the King, Book V, Chapter IX
For further reading: Antiqua et Nova (Dicastery for the Doctrine of the Faith / Dicastery for Culture and Education, January 2025) is the direct ecclesiastical precursor to the encyclical and contains more technical detail on AI governance. Rerum Novarum (Leo XIII, 1891) provides the original Social Doctrine framework on labor that Magnifica Humanitas explicitly inhabits. Gaudium et Spes (Vatican II, 1965) supplies the methodology of reading “signs of the times” that structures the encyclical’s approach.
MetaArchivist publishes long-form analysis at the intersection of records, memory, power, and the systems we build to manage all three.