Magnifica Humanitas: The Pope Writes Like the Machine He Fears

On 15 May 2026, Pope Leo XIV released Magnifica Humanitas, an encyclical letter “On Safeguarding the Human Person in the Time of Artificial Intelligence.” The document runs roughly 35,000 words across five chapters and a conclusion. It positions itself as the 135th-anniversary successor to Leo XIII’s 1891 Rerum Novarum, recasting that founding labor encyclical for the age of machine intelligence. The framing image is biblical and Manichean. Humanity is presented with a choice between two ancient construction sites. One is the Tower of Babel, where collective effort produces dominance and dehumanization. The other is the rebuilding of Jerusalem under Nehemiah, where shared responsibility under God produces communion.

Magnifica Humanitas reads with the gravitas the Vatican has spent two thousand years perfecting. The Latin title is well-chosen. The structure is meticulous. Citations to Rerum Novarum, Quadragesimo Anno, Pacem in Terris, Gaudium et Spes, Populorum Progressio, Laborem Exercens, Centesimus Annus, Caritas in Veritate, Evangelii Gaudium, Laudato Si’, and Fratelli Tutti assemble themselves into a wall of magisterial continuity. The reader is meant to feel the weight of tradition pressing against the lightness of Silicon Valley.

I am not Catholic, and the purpose of what follows is to read the document carefully against itself, paragraph by paragraph, and to show that the encyclical the Pope wrote against the dangers of artificial intelligence describes, with greater accuracy than any Vatican spokesperson would prefer, the institution that produced it. There are six places where the argument fails. The seventh and most damaging observation is structural. The encyclical is written in the prose of the thing it claims to fear.

The Babel Binary as Self-Refutation

The central rhetorical move is the choice between Babel and Jerusalem. Paragraph 9 tells the reader that technology is never neutral because it takes on the characteristics of those who design, fund, regulate, and use it. Paragraph 10 warns against what the Pope calls the Babel syndrome, defined as the idolatry of profit, a uniformity that neutralizes differences, and the pretense that a single language, even a digital one, can translate the human person into data and performance.

The objection writes itself. The encyclical condemns reductive frameworks in the same paragraph that offers one. It denounces single languages that flatten complexity in the same chapter that proposes a single biblical metaphor as the lens through which all of artificial intelligence must now be read. The form of the argument refutes its content. A Pope who warns against pretending one language can translate everything cannot then propose two ancient Hebrew construction narratives as the universal grammar for the most consequential technological transition in human history.

This is effective as a critic’s attack vector because the Pope set the trap himself. The document never pauses to ask whether the Babel frame might itself be the kind of reductive abstraction it warns against in others. It assumes the metaphor’s authority because it is biblical, which is to say because the Vatican says so. Every reader trained outside the Catholic tradition will catch this in the first ten paragraphs.

The deeper problem is that the encyclical wants the Babel reading to do too much work. Babel is meant to indict Silicon Valley, the technocratic paradigm carried over from Laudato Si’, the platform monopolies, the data brokers, and the AI labs in a single biblical gesture. The story itself was about human pride and divine intervention. The companies the Pope means to describe operate on capital flows, regulatory arbitrage, and venture math, none of which Genesis 11 contemplates. Treating these firms as a new Babel produces homily where analysis is required.

The Imago Dei as the Concealed Foundation

The encyclical claims to address all men and women of goodwill. Paragraph 2 makes the intended reach explicit. The Church wishes to engage in dialogue with all of humanity, sharing in the events, questions, and aspirations of the contemporary world.

The foundation of the argument is sectarian. At the source of social doctrine in paragraph 48 sits the mystery of the Triune God. Paragraph 50 places the imago Dei, the doctrine that humans are created in the image of the Triune God, at the heart of Christian anthropology. Paragraph 53 cites Dignitas Infinita for the proposition that every human person possesses infinite dignity grounded in being willed, created, and loved by God.

Strip out the metaphysical premise and the encyclical has nothing to say about artificial intelligence that a competent secular ethicist has not already said better. The arguments for transparency, against opaque algorithmic systems, in favor of regulatory accountability, against the concentration of digital power, in favor of attention to displaced workers, against the energy and water costs of training runs are all available in the existing literature without recourse to the Trinity. Shoshana Zuboff made the platform-power argument in 2019. Cathy O’Neil made the algorithmic-discrimination argument in 2016. Frank Pasquale made the regulatory-transparency argument in 2015. None of them required the Holy Spirit.

The Pope wants universal moral authority while reserving the metaphysical foundation to one tradition. He addresses the secular reader as if the encyclical’s principles were available on neutral ground, while the document itself insists that those principles have no ground at all without the revelation only the Church claims to possess. The reader is invited to accept the conclusions without inspecting the premises. This is the standard move of religious public reasoning, and it is the move secular philosophers from John Rawls to Jürgen Habermas have spent forty years pressing.

The encyclical does not engage that pressure. It assumes the reader will accept the bridge from creation in the image of God to dignity-bearing rights as a self-evident inference. Outside the seminaries, that bridge is the most contested move in modern moral philosophy. The document treats it as scenery.

Paragraph 89 and the Mirror

The most revealing sentence in the encyclical sits in chapter two, paragraph 89, where the Pope writes that the Church must address “spiritual, economic, institutional, sexual and power-based abuse, as well as abuses of conscience.” He treats this admission as a setup for the Church’s credibility on AI ethics. The argument is that a Church which examines itself can speak truth to the technology companies. The structure of the paragraph asks the reader to read the abuse as a wound the Church is addressing, and to grant the institution standing on that basis.

The admission is disqualifying. The encyclical demands of Silicon Valley exactly what the Vatican has never accepted for itself. Transparency about algorithms, equitable access to data, and avenues for recourse appear as demands in paragraph 71. Financial transparency has been resisted by the Vatican for centuries; the Istituto per le Opere di Religione, the Vatican Bank, remains the most opaque financial institution in Europe, the subject of multiple scandals from Roberto Calvi in 1982 through the Becciu trial in 2023. Shared responsibility across the global community for decisions that affect daily life appears as a demand in paragraph 72. The Vatican itself is governed by an absolute elective monarch whose successor is chosen by a closed conclave of cardinals he appointed. Public oversight of algorithmic decision-making and combat against opaque systems appears as a demand in paragraph 80. The doctrine of papal infallibility, defined at Vatican I in 1870, is the original unappealable algorithm.

The clergy abuse crisis is the largest institutional cover-up of the modern era. The John Jay Reports, the Boston Globe Spotlight investigation, the Australian Royal Commission, the German MHG study, the French CIASE report, and the Pennsylvania grand jury report all document the same pattern. Predators were moved between parishes. Records were destroyed. Victims were paid off or threatened into silence. Bishops who covered up were promoted. The pattern was global, sustained, and institutional. It maps directly onto what paragraph 80 condemns under the heading of algorithmic discrimination, where the algorithm at issue was Catholic Canon Law as administered by men who treated it as a shield.

A pope writing to AI engineers about transparency is writing to a mirror. The Vatican is the precise transnational private power, governed without democratic accountability, that the encyclical claims to oppose in others. The pattern I traced in my Carceral Nation work, where institutional opacity is the precondition for institutional harm, applies to the Holy See with the same force it applies to the American prison system. The encyclical’s refusal to name this resemblance is the largest piece of self-deception in the document.

Critique by Abstraction

Read chapter three, which carries the encyclical’s argument about artificial intelligence as such, and notice what is absent. The frontier labs go unmentioned. OpenAI, Anthropic, Google DeepMind, Meta, Microsoft, xAI, and Mistral are nowhere in the document. Large language models, transformer architectures, reinforcement learning from human feedback, retrieval-augmented generation: the encyclical engages none of them. Active litigation that will determine the legal foundation of the industry receives no notice, including the New York Times lawsuit against OpenAI over training data, the Anthropic settlement with authors, and the Adobe and Getty cases. The Kenyan content moderators paid roughly two dollars an hour to label child sexual abuse material so that ChatGPT could refuse to generate it do not appear. Palantir’s contracts with Immigration and Customs Enforcement go unaddressed. Deepfake pornography of underage girls now circulating in middle schools is not raised. The water that data centers in Arizona pull from the Colorado River basin and the gas turbines being installed in Memphis to power xAI’s Colossus cluster are absent from the document.

The Pope speaks at AI the way a homilist speaks at sin. The audience is meant to feel the rebuke without being asked to act. No fact on the ground in the document binds a single engineer at a single AI lab to a single specific obligation. The arguments are pitched at a level of abstraction where they cannot be falsified and cannot be implemented. The technocratic paradigm is condemned in paragraph 92 without the technocrats being named. Digital power is denounced in paragraph 95 without the powers being identified.

This is the pattern I described in The Rehearsal State. Institutional language performs concern without taking the steps concern would demand. The encyclical is a 35,000-word performance of pastoral worry that asks no specific person to do any specific thing. The Pope’s spokesman will tell reporters that the document is a moral framework, not a regulatory proposal. The spokesman is correct, and that is the criticism.

The John Paul II Problem

The encyclical leans on Saint John Paul II as the principal authority on human dignity. The pattern runs through paragraphs 51, 53, 57, 65, 66, 78, 86, and 108. Centesimus Annus is invoked on democracy, Laborem Exercens on work, Sollicitudo Rei Socialis on solidarity. The cumulative effect positions the Polish pope as the magisterial voice on the dignity of the human person in the late modern age.

Karol Wojtyła presided over the global expansion of the Catholic clergy abuse crisis. The documentation is extensive. Marcial Maciel Degollado, founder of the Legionaries of Christ, was a serial abuser of seminarians, fathered multiple children, and was protected by John Paul II until the next pope finally acted in 2006. Cardinal Bernard Law, who covered up abuse in Boston for decades, was given a sinecure at Santa Maria Maggiore by John Paul II after the Boston Globe Spotlight investigation forced his resignation in 2002. Theodore McCarrick, later laicized for abuse of seminarians and minors, was elevated to Archbishop of Washington in 2000 under John Paul II despite multiple warnings reaching the Vatican. The pattern was structural rather than incidental, and Jason Berry’s Render Unto Rome along with the documentary record at BishopAccountability lay out the institutional logic in detail.

Citing the architect of institutional silence as the voice of human dignity ranks as the encyclical’s largest unforced error. Every secular journalist who reads Magnifica Humanitas with the abuse archive open in another tab will write the same article. The Pope is calling on AI engineers to remember the dignity of every human person, with footnotes to the pope who protected the men who violated it.

The encyclical could have addressed this directly. Paragraph 89 gestures at the abuse crisis without naming the predecessor magisterium that permitted it. The gesture is insufficient. The Vatican has not yet found a way to talk about John Paul II without canonical inflation, and the consequence is that every social encyclical citing him reads, to anyone outside the curia, as a document with a missing chapter.

The Pope Already Writes Like the Machine

The deepest failure of Magnifica Humanitas lies in the prose itself, beneath any single argument the document makes. Read the encyclical without knowing who wrote it. The opening reaches for language about pivotal choices and threatened identity. In paragraph 7 the reader is instructed to discern how to navigate the era of AI. By paragraph 15 the faithful must remain profoundly human. Paragraph 22 invites the People of God to interpret history in the light of the Spirit. Here is paragraph 27 in full, near the close of chapter one, where the Pope characterizes the Church’s Social Doctrine itself:

It is not a handbook of principles and norms to be applied, but a process of shared discernment. It is born from the encounter between the eternal truth of the Gospel and the questions of history. It allows itself to be challenged by the signs of the times, and draws nourishment from the contributions of science, culture and human experience.

These are the precise tics of generated text. The block above runs anaphora on the single word “It” through three consecutive sentences. The same pattern recurs across the document on “the principle of” through ten paragraphs of chapter two. Abstract nouns stacked into prepositional cascades. Binary contrasts disguised as wisdom. Verbs of motion applied to concepts that do not move. “Navigate,” “discern,” “interpret,” “encounter.” The vocabulary belongs to a language model fine-tuned on the Vatican corpus and asked to generate one more encyclical.

I made this point in the AI-zeroing-knowledge piece earlier this year. When machine-generated abstraction reaches sufficient fluency, it becomes indistinguishable from the institutional prose it was trained on, because institutional prose was already abstract enough to be replaced. The Vatican has been generating documents in this register for two centuries. The model does not have to work hard to imitate the style. The style is a procedure the institution already produced by procedure.

We do not have to imagine an AI Pope to see the problem. The current human Pope writes AI-shaped prose about the dangers of AI, and the encyclical reads as the evidence. If Leo XIV’s successor is a language model trained on Leo XIII through Francis, the readers in the pews will not be able to tell the difference. The College of Cardinals will not notice. Even the Vatican press office will be unable to distinguish the output. What will test the faith of the faithful is the discovery that ecclesiastical and synthetic intelligence had already converged, long before any conclave bothered to look.

The Vatican has practiced its dismissal of the techno-utopian who wants to replace God with a chatbot. It has no rehearsed answer for the reader who notices that the chatbot would only have to produce what the office has been producing all along. The Pope warns against Babel. Babel is the Vatican. The encyclical describes its own author. The dignity Leo XIV wants to protect from the machines is a dignity the men in his robes have spent two thousand years administering, and the prose he wrote against the machines is the prose the machines have already learned to write.

Magnifica Humanitas is a serious document. It is well-organized, well-cited, well-translated. The pastoral concern for displaced workers, surveilled populations, and the poor is genuine and worth honoring. It is also a fossil. The institution that produced it cannot see the shape of its own reflection, and the encyclical is the proof. The future of human dignity will be argued by people who can name the companies, count the harms, and write in sentences a model cannot already produce, far from rooms with frescoes where men in white cite other men in white in other rooms with frescoes, while the actual machines watch the document upload itself to the Vatican website and learn from it for next year’s training run.