Magnifica Humanitas: Léon XIV's Encyclical on AI and Humanity

Key takeaways

1. Source. encyclical published by the Holy See, French version posted online by Bayard / La Croix on 22 May 2026. Original PDF 134 pages, signed "Given in Rome, near St. Peter's, on 15 May in the year 2026, the second of my Pontificate — LÉON PP. XIV". First official publication Friday 15 May 2026 (Rerum Novarum anniversary).
2. Author.
3. Léon XIV. (Robert Francis Prevost, born 14 September 1955 in Chicago)
4. First American pope in history (dual USA / Peru nationality)
5. Augustinian (former Prior General of the Order of St. Augustine 2001-2013)
6. Former Bishop of Chiclayo (Peru), then Prefect of the Dicastery for Bishops (2023-2025)
7. Elected 8 May 2025 — Magnifica Humanitas is his first social encyclical, ~1 year after his election
8. Strategic date. 15 May 2026 = 135th anniversary of Rerum Novarum (Léon XIII, 15 May 1891). The choice of the pontifical name "Léon XIV" plus the publication date is an explicit signal of continuity with the social tradition launched by Léon XIII on the labor question. Léon XIII had responded to the "res novae" (new matters) of the Industrial Revolution; Léon XIV responds to the "res novae" of the digital and AI revolution.
9. Structure (to remember).
10. Introduction. (n. 1-16) — Babel vs. Nehemiah, building in goodness, remaining human
11. Chapter 1. (n. 17-89) — A dynamic thought faithful to the Gospel (genealogy of the Social Doctrine from Léon XIII to Francis)
12. Chapter 2. (n. ~85-89) — Foundations and principles of the Social Doctrine (dignity, image of God, human rights + common good, universal destination of goods, subsidiarity, solidarity, social justice, integral human development)
13. Chapter 3. (n. 90-130) — Technique and mastery: technocratic paradigm + AI + transhumanism/posthumanism + grace and Christian humanism + Augustine's two cities
14. Chapter 4. (n. 131-181) — Preserving the human in the transformation: Truth, Labor, Freedom (truth as a common good, democracy, school, dignity of labor, unemployment, an economy that values dignity, family and youth, dependencies, social control, new forms of slavery)
15. Chapter 5. (n. 182-228) — Culture of power vs. civilization of love: normalization of war, force without limits, weapons and AI, crisis of multilateralism, so-called political realism, building the civilization of love (disarming words, peace through justice, the victims' perspective, healthy realism, dialogue, diplomacy)
16. Conclusion. (n. 229-245) — The Word became flesh, one body in Christ, the great work of our era, the Magnificat
17. Pivotal epistemological thesis. "the first choice is not between a 'yes' or a 'no' to technology, but between building Babel or rebuilding Jerusalem" (n. 9). The debate is not tech-vs-anti-tech; it is by what project, for whom, how.
18. Two biblical icons (structuring matrix). | Icon | Text | Symbolism | Dominant trait | |-------|-------|------------|----------------| | Tower of Babel | Gen 11:1-9 | A single, technological language, without God → confusion, dispersion | Homogenizing uniformity, profit, sacrifice of the weak, a single language (including digital) claiming to translate everything into data | | Reconstruction of Jerusalem (Nehemiah) | Neh 2-6 | Coordination of families, stone by stone, listening, prayer | Shared responsibility, diversity as a resource, listening/dialogue, God at the center, harmony of communion |
19. Critique of the technocratic paradigm. (Chap. 3):
20. Explicitly draws on Laudato si' n. 106-109 (Francis 2015) on the technocratic paradigm.
21. "The tendency to let the logic of efficiency, control and profit alone govern personal, social and economic choices" (n. 92).
22. Pivotal quotation from Romano Guardini (Das Ende der Neuzeit): "Modern man has not received the education necessary to make good use of his power" (n. 93).
23. Structural finding. "control of platforms, infrastructure, data and computing power does not belong to states, but to large economic and technological actors" (n. 95) — direct convergence with Mensch / Mistral (cf. note [mensch-mistral-commission-enquete-vulnerabilites-numeriques-souverainete-ia-2026-05-13]).
24. Magisterial definition of AI (n. 98-99 — to know by heart).
25. *"more 'cultivated' than 'constructed'". * (n. 98) — a formulation that engages directly with 2026 ML research (models emerge from an architecture + data, they are not designed line by line).
26. *"all of us, including those who design them, know little about how they actually work". * (n. 98) — the black-box thesis.
27. Categorical anti-anthropomorphism critique. (n. 99): "so-called artificial intelligences do not undergo experience, do not possess bodies, know neither joy nor pain, do not mature through relationship, do not know from the inside what love, work, friendship, responsibility mean. They have no moral conscience". AI imitates, it does not understand — "it is rather a statistical adaptation based on data and outcomes".
28. Three risks of personal AI use. (n. 100): 1. Ease of obtaining a result → fosters the habit of "over-delegating", weakens "our personal judgment and our creativity" (cf. Osmani Cognitive Surrender — note [osmani-cognitive-surrender-comprehension-debt-2026-05-05]). 2. Impression of objectivity → masks the fact that responses "reflect the cultural parameters of those who designed and trained them". 3. Simulation of human communication → "When speech is simulated, it does not build a relationship, but its appearance" — the risk that the person "loses the very desire to truly seek out the other".
29. Environmental impact explicitly named. (n. 101) — "large amounts of energy and water", "significant impact on carbon dioxide emissions", "computing power and storage capacity needs" (direct convergence with Mensch on the economics of the electron).
30. Responsibility, transparency, governance. (n. 102-108):
31. Rejection of "technical neutrality" (n. 104): "we cannot regard AI as morally neutral".
32. Risk of an algorithm that "redefines the boundaries of human possibility" (n. 103) — "the exclusion of the weakest is cloaked in neutrality and objectivity, against which it is impossible to protest".
33. Canonical position on "alignment". (n. 107): > "We cannot content ourselves with invoking the moralization of the machine, what is called the 'alignment' of AI with human values, without having the courage to add a further condition: the possibility of debating the ethical code to be used, subjecting it to shared criteria of social justice. Without this, those who control AI will impose their own moral vision, which will become the invisible infrastructure of the systems. A more moral AI is useless if that morality is decided by a handful of people." → Radical political critique of Californian AI safety without naming Anthropic or OpenAI (but evident between the lines, cf. note [wallace-wells-nyt-magazine-ai-populism-altman-backlash-no-one-ready-2026-05-08]).
34. Canonical concept of "disarming AI". (n. 110, to remember): > "Disarming AI means removing it from the logic of armed competition, which today is no longer only military but also economic and cognitive. It is the race for the highest-performing algorithm and the largest database, aimed at consolidating a geopolitical or commercial advantage over everyone else. To disarm is to break this equivalence between technical power and the right to govern. [...] It means removing it from monopolies, making it debatable, contestable, and thus habitable, by restoring it to the plurality of human cultures and ways of life."
35. Special appeal to AI developers. (n. 111) — "Technological innovation can be, in a certain sense, a human form of participation in the divine act of creation. Developers therefore bear a particular ethical and spiritual responsibility, since every design choice expresses a vision of humanity." No stigmatization, but a theological elevation of the developer's role, coupled with a call for greater responsibility.
36. Critique of transhumanism and posthumanism. (n. 115-117) — "an archipelago of different conceptual islands, nonetheless linked by the same ocean of assumptions: the centrality of technique and the dream of surpassing the limits of the human condition" (n. 116). Explicit distinction:
37. Transhumanism. = enhancement of the human through technologies (biomedicine, bodily engineering, devices, algorithms), increasing performance/capabilities.
38. Posthumanism. = critique of anthropocentrism + human/machine/environment hybridization, "crossing a threshold where humanity will surpass itself by entering a new evolutionary stage".
39. Categorical rejection. (n. 117): "if the human being is treated as material to be perfected or surpassed, it then becomes easier to accept that some be considered less useful, less desirable, less worthy. In the name of progress, one may come to imagine 'necessary sacrifices' and to make the most fragile pay the price of a supposed optimization of the species."
40. The true "more than human". (n. 127-128):
41. Not transhumanism. , but grace: "we manage to be fully human when we are more than human, when we allow God to lead us beyond ourselves so that we may attain our truest being" (n. 128, citing Francis, Evangelii gaudium n. 8).
42. Quotation from St. Thomas Aquinas on the infinite disproportion between nature and divine life, bridged by God himself.
43. Pivotal epistemological distinction. "For an algorithm, error is something to correct; for a person, it can be the start of profound change" (n. 128).
44. Augustine's two cities. (n. 130) — the theological matrix: "Two loves made two cities: love of self even to contempt of God, the earthly city; love of God even to contempt of self, the heavenly city" (Augustine, De civitate Dei XIV, 28). "The age of AI is no exception to this rule: the building of Babel or of Jerusalem begins within each of us."
45. Canonical question drawn from John Paul II's *Redemptor hominis. (1979) — n. 129: "Does AI make human life on earth 'more human' in every respect? Does it make it more 'worthy of man'?"* — the ultimate criterion for evaluating AI according to Léon XIV.
46. Truth as a common good. (Chap. 4, n. 132-147):
47. AI "accelerates the profound changes affecting public and political communication" (n. 132).
48. "Disinformation was not born with AI, but today finds in it a powerful multiplier" (n. 132).
49. "Truthful information does not arise from centralized or automated control" (n. 132) — rejection of a "ministry of truth".
50. Truth as a relational good. "truth is even more relational; it is built through bonds of trust and shared practices".
51. Dignity of labor in the digital transition. (n. 148-164):
52. Central quotation. , Antiqua et nova n. 67 (n. 150): "while AI promises to boost productivity [...] workers are often forced to adapt to the speed and demands of machines, instead of the machines being designed to help those who work. Thus, contrary to the advertised benefits of AI, current approaches to technology can paradoxically deskill workers, subject them to automated surveillance, and relegate them to rigid and repetitive tasks."
53. Explicit concern. "it is realistic to fear a significant and rapid contraction in available jobs, with a domino effect that deeply affects families, young people and local economies" (n. 151).
54. Paradox of material progress / anthropological regression. (n. 154) if too few people work.
55. Four policy axes. (n. 156): social criteria for innovation + active training/retraining policies + corporate responsibility + dignity of labor as an indicator of success.
56. GDP obsolete. (n. 159) — call for alternative indicators.
57. Implicit convergence. with Andrew Ng's Jobpocalypse (note [ng-the-batch-352-no-ai-jobpocalypse-2026-05-08]), Mollick × roon ASI (note [mollick-roon-asi-consulting-forward-deployed-engineering-2026-05-10]), Sun NYT Permanent Underclass (note [sun-nyt-silicon-valley-permanent-underclass-2026-04-30]).
58. New forms of slavery. (n. 170-181 — a very forceful section):
59. Digital dependencies. (n. 170): "digital attention economy", "platforms and services designed to capture users' time and attention".
60. Social control through profiling. (n. 171): "architecture of visibility", "what is amplified or made invisible" → "conformism and self-censorship".
61. Technocratic-posthumanist mentality. (n. 172): "some posthumanist currents even go so far as to envisage 'second-class' human beings".
62. Invisible labor of microworkers. (n. 173, a crucial denunciatory section to know by heart): > "A significant part of how the digital economy functions rests on the silent labor of millions of human beings, employed in activities that are barely visible but essential: data labeling, content moderation — often of a very disturbing kind — and model training. In many cases these are young people, mostly women, who labor for a pittance. To this invisible fatigue is added the even more brutal fatigue of extracting the resources needed to produce the devices and microprocessors on which AI depends. In certain regions of the world, children and adolescents work in dangerous conditions crushing the materials from which rare earths are extracted. Bodies marked, mutilated, used so that the flow of computation never stops."
63. Historic request for forgiveness. on slavery (n. 176): "in the name of the Church, I sincerely ask for forgiveness" — a strong gesture following John Paul II.
64. Data colonialism. (n. 178, canonical concept): > "Colonialism today wears an unprecedented face. It no longer dominates only bodies, but appropriates data, turning personal lives into exploitable information. [...] These are the new 'rare earths' of power: vital information that, once cross-referenced, can be used to train predictive models, guide investment strategies, anticipate crises and, above all, select which people and things matter. [...] Otherwise, the digital era will not be postcolonial, but colonial in a new form."
65. Three levers for action. (n. 179): transparency of production chains + ethical due diligence by companies + platform cooperation with authorities.
66. Weapons and AI. (Chapter 5, n. 197-200 — a section critical of military AI):
67. Categorical rejection. of artificial moral agents (n. 198): "We sometimes speak of 'artificial moral agents' as if a machine could guarantee, more consistently than a human being, the distinction between good and evil. But moral judgment cannot be reduced to a mere calculation. [...] It is therefore not acceptable to entrust artificial systems with lethal or, in any case, irreversible decisions. No algorithm capable of making war morally acceptable."
68. Three criteria for discernment. (n. 199): 1. Traceable personal responsibility — "the chain of responsibility must remain identifiable and verifiable". 2. Time for moral judgment — "AI tends to shorten decision times; but in wartime, irreversible decisions cannot have speed and efficiency as their supreme criteria". 3. Identification and protection of civilians — "Any technology that makes it easier to strike without seeing the other's face lowers the moral threshold of conflict."
69. Requirements. (n. 200): traceability, effective human control over lethal decisions (a ban on autonomous weapons), shared international rules.
70. Convergence with AISI UK / Mythos. (note [aisi-uk-gpt55-cyber-capabilities-evaluation-2026-04-30]) on emerging cyber capabilities.
71. Five paths toward a civilization of love. (n. 213-227): 1. Disarming words — "Let us disarm words and we will help disarm the Earth" (n. 214) 2. Building peace through justice (n. 215) 3. Adopting the victims' perspective (n. 216-217) 4. Cultivating a healthy realism (n. 218) — neither political idealism nor cynicism/Realpolitik 5. Reviving dialogue (n. 219-227) — interpersonal + interreligious + diplomatic + multilateralism
72. Quotation from J.R.R. Tolkien (Sam, in The Return of the King). (n. 213): > "It's not our part to master all the tides of the world, but to do what is in us for the succour 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." A surprising literary reference (Tolkien was a practicing Catholic) that anchors the document in twentieth-century Western popular culture.
73. Conclusion: a four-part spiritual program. (n. 229-245):
74. Faith. (The Word became flesh) — contemplating the Father's design
75. Charity. (One body in Christ) — Eucharist + ecclesial unity
76. Hope. (The great work of our era) — the wise architect + Nehemiah
77. Prayer. (The song of the Magnificat) — Mary as poet and prophetess of redemption
78. Main magisterial sources to know. | Document | Date | Relevance | |----------|------|------------| | Rerum Novarum (Léon XIII) | 15 May 1891 | 135th anniversary — the original matrix of the Social Doctrine | | Quadragesimo anno (Pius XI) | 1931 | Principle of subsidiarity, critique of economic power concentration | | Mater et Magistra + Pacem in terris (John XXIII) | 1961, 1963 | Extension to human rights + universal address | | Gaudium et spes (Vatican II) | 7 December 1965 (60th anniversary in 2025) | Church's presence in the world + autonomy of temporal realities | | Populorum progressio (Paul VI) | 1967 | Integral development, link between peace and development | | Octogesima adveniens (Paul VI) | 1971 | No single response valid for all contexts | | Laborem exercens (John Paul II) | 1981 | Labor as the essential key to the social question | | Sollicitudo rei socialis + Centesimus annus (John Paul II) | 1987, 1991 | Permanent paradigm of the Social Doctrine | | Redemptor hominis (John Paul II) | 1979 | Canonical criterion: does technique make life more worthy? | | Caritas in veritate (Benedict XVI) | 2009 | Finance + technology + ethical integration | | Evangelii gaudium (Francis) | 2013 | Time is greater than space + the image of the polyhedron | | Laudato si' (Francis) | 2015 | Technocratic paradigm + our common home | | Fratelli tutti (Francis) | 2020 | Civilization of love + moving beyond "just war" | | Dilexit nos (Francis) | 2024 | Spirituality of the heart | | Antiqua et nova (DDF + DCE) | 14 January 2025 | Joint note on AI — the most-cited upcoming magisterial source | | Quo vadis, humanitas ? (ITC) | 9 February 2026 | Christian anthropological reflection on scenarios for humanity's future |
79. Connections within the watch dossier.
80. Direct convergence with Mensch / Mistral. (note [mensch-mistral-commission-enquete-vulnerabilites-numeriques-souverainete-ia-2026-05-13]): Léon XIV takes up the concentration of technological power in the hands of private actors (n. 5, 95) — "Today the main drivers of development are private actors, often transnational, endowed with resources and capacities for action greater than those of many governments". But where Mensch speaks of economic sovereignty, Léon XIV speaks of ethical-anthropological sovereignty.
81. Convergence with Mollick × roon ASI. (note [mollick-roon-asi-consulting-forward-deployed-engineering-2026-05-10]): recognition that AI transformation is organizationally slow and complex (a Hayekian problem).
82. Convergence with Andrew Ng's *Jobpocalypse. (note [ng-the-batch-352-no-ai-jobpocalypse-2026-05-08]): Léon XIV acknowledges both that "AI is disrupting work" AND that "there is no single model of change"* (n. 153).
83. Convergence with Wallace-Wells's AI Populism. (note [wallace-wells-nyt-magazine-ai-populism-altman-backlash-no-one-ready-2026-05-08]): implicit critique of Californian tech oligarchs, rejection of the Californian "safety-first" discourse (n. 107 on alignment).
84. Convergence with Sun's NYT Permanent Underclass. (note [sun-nyt-silicon-valley-permanent-underclass-2026-04-30]): the labor → capital shift, invisible microworkers.
85. Convergence with Osmani's Cognitive Surrender. (note [osmani-cognitive-surrender-comprehension-debt-2026-05-05]): Léon XIV n. 100 on the "ease of obtaining a result" that "weakens our personal judgment and our creativity".
86. Convergence with AISI UK's GPT-5.5 cyber. (note [aisi-uk-gpt55-cyber-capabilities-evaluation-2026-04-30]) on emerging offensive capabilities.
87. Convergence with Uber Engineering's Agent Identity. (note [uber-engineering-agent-identity-crisis-zero-trust-spire-2026-05-21]) on the need for a "chain of responsibility that is identifiable and verifiable" (n. 199) — where Uber provides a technical answer, Léon XIV provides an ethical one.
88. Convergence with Bain's $100B SaaS Cross-System Labor. (note [bain-100b-saas-opportunity-cross-system-labor-agentic-ai-2026-05]) on AI vendors' pricing power anchored to wages.
89. Convergence with Bain's Rule of 40 SaaS. (note [bain-ai-rule-of-40-headwinds-tailwinds-saas-2026-04]) on the transformation of business models.
90. Convergence with DORA × Google Cloud's J-Curve. (note [dora-google-cloud-roi-ai-assisted-software-development-j-curve-2026-04-21]): Léon XIV insists on "the quality and dignity of labor among the indicators of success" (n. 156) — resonating with the DORA report's "Code is a liability, not an asset."
91. Weaknesses / open questions.
92. Absence of concrete figures. (unlike parliamentary audits such as Mensch) — a purely doctrinal register.
93. No naming. of tech companies (OpenAI, Anthropic, Google, Meta, Microsoft, etc.) — a deliberate diplomatic choice that may limit the critique's reach.
94. Position on open vs. closed models. not made explicit — even though the question is central to sovereignty.
95. The relationship between the Church and AI regulation (EU AI Act, Bletchley, etc.). is not specified — the document sets out principles but does not engage with the details of ongoing legal frameworks.
96. The relationship to the attention-economy business model. is addressed (n. 170) but not developed in depth regarding specific mechanisms.
97. The critique of autonomous weapons. is forceful (n. 197-200) but there is no explicit call for an international moratorium as demanded by some governments and NGOs.
98. Léon XIV's key vocabulary to remember. magnificent humanity, Babel vs. Jerusalem, cultivated not constructed AI, disarming AI, disarming words, new AI monopolies, epistemic asymmetry, data colonialism, new rare earths of power, invisible labor, artificial moral agents (rejection), one body in Christ, wise architect, Nehemiah as parable, Magnificat as poet of redemption, invisible infrastructure of systems (on alignment), situated anthropocentrism.
99. Useful for.
100. Executive presentations on the "social and ethical responsibility of AI" in business (a non-technical canonical reference with broad cultural grounding).
101. EU AI policy / multilateralism discussions (Brussels, UN AI Advisory Body, Bletchley, etc.).
102. Executive/manager training on AI ethics (continuing-education modules at ESCP, HEC, INSEAD, IFG).
103. Argument in favor of democratic regulation of AI (cf. n. 107 on alignment).
104. A dossier on "new forms of digital slavery" (data labeling, moderation, rare earths) for CIOs / CSR teams.
105. A reference in any document on the dignity of labor in the age of automation.
106. Comparison of Catholic Social Doctrine vs. other AI ethics doctrines (Asilomar, OECD, UNESCO, IEEE).
107. A dossier on "ethical sovereignty vs. technical sovereignty" (the junction of Léon XIV ↔ Mensch).
108. Argument material for homilies / catechesis / seminarian training.
109. Board-of-directors briefings on AI and fiduciary responsibility.