# From the Vatican to Silicon Valley: the encyclical that turned AI into a moral dispute about power

Author: Fernando Nieto Lobato
Original publication: 2026-06-03
Spanish original: https://estrategiabyaleph.substack.com/p/estrategia-140-desarmar-la-ia-el
English URL: https://elcontemplador.github.io/estrategia-english/essays/140/
Status: Published translation

This is a translation of the original Spanish essay published on 3 June 2026. Its claims, examples and forecasts retain that historical context.

English publication: 2026-09-29

*Editorial introduction:* Dear readers, we know that, in a sense, we are “late” in analysing Leo XIV’s latest encyclical on artificial intelligence. [Last week we already had an interesting guest article scheduled](https://open.substack.com/pub/estrategiabyaleph/p/estrategia-139-la-nueva-fase-de-la?r=2tyybn&utm_campaign=post-expanded-share&utm_medium=web), and since [Magnifica humanitas](https://www.vatican.va/content/leo-xiv/es/encyclicals/documents/20260515-magnifica-humanitas.html) was publicly presented on Monday 25 May, it has generated an avalanche of articles, opinion pieces, commentaries and reactions across general, technological, religious and political media. Having now had the chance to read it calmly this week — it is highly recommended reading, addressing many subjects beyond AI — we do not think it makes sense to offer a simple first reading as though the debate had not already begun.



So in this issue we have preferred to do something slightly different. Rather than merely summarising or commenting on the encyclical, although we will inevitably have to identify some of its central points, we will observe the debate it has prompted. We are interested both in the document and in what has happened around it: favourable readings, criticism, discomfort, political appropriations, reservations in the technology world, Anthropic’s presence at the presentation, and the way a papal text has forced very different actors to take a position on the same question: who should provide moral direction for artificial intelligence?



Before leaving you with the article, it is worth remembering that this is not the first time estrategIA has considered the Church’s reflections on artificial intelligence: [in issue 81 we analysed Antiqua et Nova, the Vatican document that already placed human dignity, truth and ethical responsibility at the centre of the technological debate](https://open.substack.com/pub/estrategiabyaleph/p/estrategia-81-la-tradicion-interpela?r=gp8ei&utm_campaign=post-expanded-share&utm_medium=web). In a sense, Magnifica humanitas can be read as another step in the same conversation, now raised to the highest symbolic level of an encyclical.



*Translation note: this article preserves the debate as reported on 3 June 2026. Quotations from the encyclical, Tolkien and commentary are translated from the Spanish source unless explicitly identified as English visible in a screenshot; they are not presented as verified official English wording.*



The real news is not that a pope has warned about the risks of artificial intelligence. Governments, universities, laboratories, trade unions, international organisations, activists, philosophers, scientists and regulators have already done that. What is new is the Church’s decision to place AI at the heart of a new social question, comparable — at least in symbolic ambition — to the one [Rerum novarum](https://www.vatican.va/content/leo-xiii/es/encyclicals/documents/hf_l-xiii_enc_15051891_rerum-novarum.html) posed in response to the Industrial Revolution in 1891.



Then, Leo XIII looked towards factories, wages, workers’ conditions and the concentration of industrial capital. Now, Leo XIV looks towards data, algorithms, platforms, computing power, automation and the concentration of technological power. The underlying question is similar: what happens to the person when a new economic and technical infrastructure reorganises work, social life and the distribution of power?



Initial international coverage, from [Reuters](https://www.reuters.com/business/media-telecom/pope-leo-urges-world-slow-down-ai-fervent-first-manifesto-2026-05-25/) to [TIME](https://time.com/article/2026/05/25/pope-leo-encyclical-ai-magnifica-humanitas/), immediately placed the document at that intersection of regulation, work, war and technological power. [Vatican News](https://www.vaticannews.va/en/pope/news/2026-05/pope-leo-xiv-encyclical-magnifica-humanitas-ai.html) captured its core in a clear formulation: **AI must serve humanity, not concentrate more power in the hands of a few**. That is why the impact of Magnifica humanitas cannot be understood through its chapters alone. It must also be understood through the reactions. The text acts as a mirror: every actor commenting on it reveals something of their own position on artificial intelligence.



## AI is not neutral



One point repeatedly highlighted in early readings of the encyclical is its outright rejection of technological neutrality. For Leo XIV, artificial intelligence is not a toolbox that is subsequently used well or badly. Its design incorporates priorities, interests, worldviews and power relations. A system that classifies, predicts, recommends, monitors, evaluates or automates decisions does more than process information: it participates in organising social life.



This connects with an increasingly widespread intuition: AI cannot be analysed solely through its technical performance. It is not enough to ask whether a model is faster, cheaper, more multimodal or more capable. We must ask who trains it, with what data, under which incentives, with what oversight and business model, and with what effects on work, democracy and personal autonomy.



Here lies one of the document’s strongest ideas: a “more moral” AI is not enough if that morality is defined by a handful of people. The criticism directly targets the familiar discourse of “alignment” with human values. The decisive question is not simply how to align models, but who decides those values, with what legitimacy, according to which criteria of justice and with what scope for democratic challenge. Programmer and entrepreneur [Simon Willison](https://simonwillison.net/2026/May/25/encyclical-on-ai/) has praised precisely that ethical clarity.



## “Disarming AI”



The phrase attracting most attention has been “disarming AI”. It is a powerful formulation, perhaps deliberately uncomfortable. It works because it condenses several dimensions of the debate: militarisation, geopolitical competition, the corporate race, concentration of computing power, surveillance, information manipulation and the temptation to turn algorithmic systems into instruments of domination.



At the official presentation, reported by [Vatican News](https://www.vaticannews.va/en/pope/news/2026-05/pope-leo-xiv-magnifica-humanitas-presentation-ai-disarmament.html), Leo XIV linked the verb to the need to free AI from the logics of domination, exclusion and war. The metaphor has risks: it may sound like technophobia or an indiscriminate brake on innovation. Read carefully, however, it does not mean switching off data centres or abandoning beneficial uses. It means subjecting algorithmic power to the rule of law, democratic deliberation, public responsibility and verifiable ethical limits.



The military dimension is obvious. The encyclical warns against autonomous weapons and delegating lethal decisions to machines. In a world where AI already plays a role in military intelligence, target selection, drone warfare, cyberdefence and strategic surveillance, that warning has uncomfortable implications. The [Financial Times](https://www.ft.com/content/09089fa8-d2d8-4328-92b5-5a4a186e6523) has pointed to precisely this difficulty: **the logic of the international AI race makes it very hard for a moral appeal, however powerful, to change governments’ and companies’ incentives on its own**.



But “disarmament” goes beyond the military. It also targets the civilian weaponisation of AI: systems that manipulate attention, determine public visibility, make work more precarious, profile behaviour or shape democratic processes. [TechCrunch](https://techcrunch.com/2026/05/25/the-popes-ai-encyclical-isnt-really-about-ai/), in one of the readings with which we feel most in tune, put it suggestively: **perhaps the encyclical is not really about AI, but about older problems that AI intensifies — inequality, war, democratic erosion and the concentration of power**. AI does not create our age’s tensions from nothing, but it can turn them into infrastructure.



This reading also makes Europe uncomfortable. [EU News](https://www.eunews.it/en/2026/05/26/the-eu-responds-to-the-popes-call-on-ai-an-effective-legal-framework-is-already-in-place-in-europe/) connected the encyclical with the AI Act and the European Commission’s response, which said it shared the Pope’s diagnosis. But **the debate itself recalls a limit: Europe’s most ambitious rules coexist with sensitive exceptions for national security and defence. It is precisely there, where states most resist surrendering sovereignty, that the verb “disarm” becomes more politically demanding**.



## Anthropic in the Synod Hall



One of the most discussed elements of the presentation was the presence of Christopher “Chris” Olah, an Anthropic co-founder and a prominent figure in interpretability research. Presenting a papal encyclical on AI alongside a representative of one of the most important frontier laboratories turns the event into a symbolic scene: the Vatican and Silicon Valley sharing a stage to discuss the moral limits of a technology reorganising the world.



The most generous reading is obvious: the Church does not want to discuss AI from outside, without listening to those building it; and Anthropic has spent years trying to distinguish itself as a company concerned with safety, interpretability and the risks of advanced AI. In his [remarks, published by Anthropic](https://www.anthropic.com/news/chris-olah-pope-leo-encyclical), **Olah raised three issues: job displacement, the global distribution of benefits, and the opacity of increasingly incomprehensible models. He also made an unusual admission: frontier AI laboratories operate under commercial, geopolitical and personal incentives that can conflict with doing the right thing.**



But a critical reading is necessary too. [Wired](https://www.wired.com/story/anthropic-christopher-olah-pope-ai-encyclical/) interpreted the scene as an unprecedented alliance between the Church and Silicon Valley; [National Catholic Reporter](https://www.ncronline.org/vatican/vatican-news/pope-leo-anthropic-co-founder-call-church-tech-ethics-partnership-magnifica) and [OSV News](https://www.osvnews.com/anthropics-christopher-olah-urges-global-moral-oversight-of-ai-at-vatican-presentation/) highlighted Olah’s call for global moral oversight. Anthropic gains a great deal from that image: it reinforces its narrative as a responsible laboratory and accrues reputational capital. The uncomfortable question is unavoidable: where does necessary dialogue between moral institutions and technology companies end, and ethics-washing begin?



Perhaps the most honest answer is to accept the ambivalence. Dialogue is necessary; uncritical legitimation would be dangerous. The Church needs to understand the real technology, not a caricature. Companies need to listen to institutions that are not subject to the same incentives. But no company should become a privileged interpreter of the common good.



## An anecdote very much of our time



Even authorship has become part of the debate. [The Verge](https://www.theverge.com/ai-artificial-intelligence/937801/pope-leo-xiv-magnifica-humanitas-ai-pangram) reported that the Pangram detector identified possible AI authorship in passages of Magnifica humanitas. This needs great caution: **these detectors make mistakes**, especially with translated, formulaic or highly institutional writing. But the anecdote is revealing. In an encyclical about protecting the human in the age of artificial intelligence, even the suspicion of AI-assisted writing becomes a metaphor for our times.



[![The Verge headline about possible AI authorship of the encyclical, with date, authors and Getty Images credit.](https://elcontemplador.github.io/estrategia-english/assets/images/d818f54e-8b9d-443c-bc2c-68b0da2d5b0b_1044x1148.png)](https://substackcdn.com/image/fetch/$s_!u8Gy!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd818f54e-8b9d-443c-bc2c-68b0da2d5b0b_1044x1148.png)



*Transcription of the English headline and publication details visible in the screenshot:* The Verge, AI/News. **“Did the Pope use AI to write about the dangers of AI?”** Subheading: “Analyses have determined that parts of the Magnifica Humanitas appear to have been written by AI.” By Jay Peters and Hayden Field. 27 May 2026, 2:38 a.m. GMT+2; 67 comments. The photograph shows the Pope speaking at a microphone and is credited to Getty Images. This is the outlet’s displayed claim; the article above expressly cautions against treating detector output as proof of authorship.



## Reactions: enthusiasm, caution and dispute



The reactions reveal a telling map of interpretations. **Across the Catholic world, favourable reception has been broad. Many have read the text as a major new social encyclical: not a condemnation of AI, but a defence of the human person against a revolution threatening to reduce them to productivity, data or a predictive profile.** [Paolo Carozza](https://law.nd.edu/news-events/news/paolo-carozza-provides-insight-on-pope-leo-xivs-encyclical-magnifica-humanitas/), a professor at the University of Notre Dame, placed it in continuity with Catholic social teaching; [America Magazine](https://www.americamagazine.org/short-take/2026/05/26/magnifica-humanitas-ai-pope-leo-human-dignity/) interpreted it as an explicit defence of human dignity in the age of AI.



**The technology world’s reaction has been more divided.** Some voices welcomed the fact that the Pope does not obsess over futuristic superintelligence scenarios but addresses present harms: monopolies, disinformation, precarious work, autonomous weapons, inequality and democratic fragility. **Others criticise him for not taking the possibility of AGI seriously enough. [The Verge](https://www.theverge.com/ai-artificial-intelligence/937933/pope-ai-encyclical-tech-industry-reactions) summed this up with an expressive headline: the Pope is “not AGI-pilled”. Many technologists wanted an encyclical about where AI is going; Leo XIV has instead written one about what AI is already doing to us.**



**There have also been selective political appropriations.** Pedro Sánchez linked the encyclical to an AI that is not neutral and is oriented towards the common good, as reported, for example, by [La Voz de Galicia](https://www.lavozdegalicia.es/noticia/sociedad/2026/05/26/sanchez-alinea-papa-leon-xiv-frente-ia-neutral-favor-paz/00031779786643253200967.htm). [Claudia Sheinbaum](https://www.infobae.com/mexico/2026/05/26/enciclica-de-leon-xiv-es-aplaudida-por-sheinbaum-quien-detenta-la-inteligencia-artificial-es-muy-interesante/) highlighted the question of who controls artificial intelligence, while [Gustavo Petro](https://www.infobae.com/colombia/2026/05/26/petro-compartio-el-magnifica-humanitas-del-papa-leon-xiv-y-advirtio-sobre-la-inteligencia-artificial-despedira-centenares-de-trabajadores/) shared the document to warn about its employment effects. That is understandable: governments need strong moral language to justify interventions the market resists. But there is also a risk of cherry-picking: embracing criticism of technological power while ignoring the less comfortable anthropological passages.



## Babel, Tolkien and the Ring



**Perhaps the encyclical’s deepest contribution is its criticism of a widespread idea: that what is most efficient is, by definition, most desirable.** AI arrives wrapped in promises of optimisation: optimise work, education, administration, health, security, communication and citizen services. But optimising does not always mean humanising. Sometimes it means accelerating processes without asking who is left out, which capacities atrophy, which bonds break or which responsibilities become diluted.



**In our view, one of the most suggestive readings comes from [Daniel Arjona in El Arjonauta](https://elarjonauta.substack.com/p/el-papa-contra-la-ia-magnificas-humanitas)**. He interprets Magnifica humanitas as an encyclical threaded through with two biblical images: Babel, symbolising the ambition to translate everything — including the mystery of the person — into data and performance; and Jerusalem, rebuilt stone by stone by Nehemiah, as an image of a humanity that does not give up on building a shared future.



Arjona also highlights a beautiful detail which, as a member — albeit not a very active one — of the Spanish Tolkien Society, I cannot fail to mention: **the encyclical quotes Tolkien. Not as literary decoration, but as a moral key within the chapter devoted broadly to “war”. Against the temptation of the Ring — a single power “to rule them all” — Leo XIV revives Tolkien’s intuition of doing “what is in our hands” in the time we have been given to live.** Gandalf’s words fit AI almost too well: this is not only about building more powerful systems, but about preventing that power from being concentrated in a few hands capable of governing other people’s lives through an invisible infrastructure.



[![Spanish paragraph 213 of the encyclical, quoting Tolkien and setting out five paths of responsibility.](https://elcontemplador.github.io/estrategia-english/assets/images/91ce9f79-bfa5-4552-826d-f8954a16fe0d_1124x444.png)](https://substackcdn.com/image/fetch/$s_!9uGH!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F91ce9f79-bfa5-4552-826d-f8954a16fe0d_1124x444.png)



*English translation of the Spanish excerpt visible in the image, paragraph 213 of the encyclical. The embedded Tolkien passage is also translated from that Spanish text; this is not a transcription of Tolkien’s published English or of an official English version of the encyclical.*



> 213. A twentieth-century Catholic writer, John Ronald Reuel Tolkien, described our responsibility through the words of one of the protagonists in one of his novels: “It is not for us to master all the tides of the world, but to do what is in our hands for the good of the days we have been given to live, rooting out evil in the fields we know, and leaving those who come after us a clean land to till.” [187] The civilisation of love is born not of a single spectacular gesture, but of a sum of small, tenacious acts of faithfulness that confront dehumanisation. It is therefore worth pausing to consider some ways in which each of us, within our own sphere, can help build it. Without seeking to exhaust the subject, I propose five paths of everyday and public responsibility: disarming words, building peace in justice, adopting the perspective of victims, cultivating a healthy realism, and relaunching dialogue and multilateralism.



This reading has not stood alone. [Jorge Morla in El País](https://elpais.com/cultura/2026-05-27/la-lucha-entre-el-papa-y-los-tecnooligarcas-a-traves-de-el-senor-de-los-anillos.html) connected the Tolkien quotation with Palantir, the company named after Middle-earth’s seeing stones which symbolises, like few others, the convergence of data, surveillance and power. [James T. Keane in America Magazine](https://www.americamagazine.org/catholic-book-club/2026/05/26/why-pope-leos-new-encyclical-quotes-gandalf-literary-images-of-hope-and-faith-in-magnifica-humanitas/) explained why Gandalf works as an image of hope and responsibility, while [Wired](https://www.wired.com/story/pope-leo-schooled-the-tech-bros-on-tolkien) read the reference as a correction to technological appropriations of Tolkien by parts of Silicon Valley.



Here lies a powerful cultural key. For years, Silicon Valley has built an almost religious mythology around its own mission: save the world, accelerate progress, transcend limits, redesign the human condition. The encyclical responds with another set of images: not the single power that sees and governs everything, but the patient, plural and communal rebuilding of a habitable city. [The New Yorker](https://www.newyorker.com/news/the-lede/what-pope-leo-xiv-said-about-ai) read the text precisely as a critique of Silicon Valley’s cultural imagination and its tendency to present AI in almost salvific language.



## A compass, not a final answer



The encyclical does not solve every problem. It does not go into sufficient detail on some technical debates. It does not offer a complete regulatory architecture. Its symbolic relationship with Anthropic merits caution. And some of its anthropological premises will not be shared by many people who nevertheless may agree with much of its social diagnosis.



But perhaps that was not its function. **Magnifica humanitas does not seek to be an engineering manual or a European regulation. It seeks something prior to that: to change the frame of the conversation. To remind us that artificial intelligence is not only a race to build more capable systems, but a dispute about what we mean by humanity, who concentrates power, how truth is protected, what value we place on work, what limits we set on war and which institutions should safeguard the common good.**



The underlying question is not whether the Church has the last word on AI. It does not. Nor do technology companies, states, experts, religious communities or markets. That is precisely why the debate matters, and why we have spent 140 issues of this newsletter trying to encourage it.



Artificial intelligence is forcing societies to decide which parts of human life should be automated, which protected, which shared, and which must never become objects of calculation or business. Magnifica humanitas does not close that conversation. It opens it with a symbolic force that is difficult to ignore. Perhaps that is its most valuable contribution: reminding us that AI’s future will not be only a question of technical capability, but of moral imagination, political responsibility and institutional courage.



[Fernando Nieto Lobato](https://www.fernandonieto.es/bio/)



*Director of Digital Innovation at Institución Educativa ALEPH and editor of the estrategIA newsletter*



[![estrategIA infographic on eight themes in the debate around Magnifica humanitas.](https://elcontemplador.github.io/estrategia-english/assets/images/c3b14b85-0a85-4aca-86b1-a8f8f52846fd_1254x1254.png)](https://substackcdn.com/image/fetch/$s_!A7BH!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc3b14b85-0a85-4aca-86b1-a8f8f52846fd_1254x1254.png)



*English transcription of the accompanying infographic: “Magnifica humanitas: the debate opened by the encyclical on AI — Key themes shaping the global conversation”. At its centre: “From the Vatican to Silicon Valley: a moral dispute about technological power”.*



1. **A new social question.** AI stands at the centre of a debate comparable to that of the Industrial Revolution: what happens to the person when technological power reorganises work, society and the distribution of power?



2. **AI is not neutral.** Designs incorporate interests, priorities and worldviews. The key question is not only how to align models, but who decides the values, with what legitimacy and under what controls.



3. **“Disarming AI”.** Free AI from the logics of domination, militarisation, surveillance, manipulation and concentrated power. Subject algorithmic power to the rule of law, democratic deliberation and verifiable ethical limits.



4. **Who decides the values?** At the centre is the legitimacy of those defining AI ethics. Risk of “private alignment” without public participation or the possibility of democratic challenge.



5. **Uncomfortable alliances.** Dialogue between moral institutions and AI laboratories is necessary but ambivalent. Listen to those building the technology, yes; grant them moral legitimacy without criticism, no.



6. **Diverse reactions.** Positive reception in many social and religious circles. Division in technology: some welcome the focus on present harms, while others miss greater emphasis on future risks. Selective political appropriations also occur.



7. **Competing cultural visions.** Between the “Ring” of power that sees and governs everything and the patient construction of a shared, plural and human future. This is not about possessing greater capability, but about which humanity we want to protect.



8. **A compass, not a final answer.** The encyclical offers neither a technical nor a regulatory manual. Its strength lies in reframing the debate: reminding us that AI’s future will, above all, be a question of moral imagination, political responsibility and institutional courage.



The five closing principles are human dignity, the common good, public responsibility, democracy and participation, and ethical limits. Footer, estrategIA #140: “The decisive question is not whether AI can do more. It is who should decide what for, and for whom.”