Pope Leo XIV

Who he is

Affiliation: 267th Bishop of Rome, elected 2025-05-08 (b. Robert Francis Prevost, Chicago, 1955-09-14 — turned 71 the week before this review). Status check run 2026-09-21: alive, in office, publishing weekly (Angelus 2026-09-20; audience with the Leo XIV International Study Center 2026-09-18), four days out from a UNESCO keynote on AI in Paris (France trip 25–28 Sept 2026). First American pope; naturalized Peruvian (2015, per Wikipedia — no primary fetched); Augustinian friar (O.S.A.). Career, re-verified this cycle against the Vatican News biography: BS mathematics, Villanova (1977); M.Div., Catholic Theological Union (Chicago); licentiate (1984) and doctorate (1987) in canon law at the Angelicum, Rome — thesis The Role of the Local Prior in the Order of Saint Augustine; Peru missions at Chulucanas (1985–86) and Trujillo (1988–99, seminary prior, canon-law professor, judicial vicar); Prior General of the Augustinians (elected 2001, second term 2007, to 2013); Apostolic Administrator (2014-11-03) then Bishop of Chiclayo (2015-09-26 to 2023); Prefect of the Dicastery for Bishops (2023-01-30 to election); cardinal 2023-09-30. Two v1 facts corrected: the doctorate is from the Angelicum, not "Catholic University of America" (v1 carried CUA behind a [TBD]); Peru service was two stints, not a continuous "1985–1998".
One-line position: AI is this century's Rerum Novarum moment, and the first encyclical of his papacy — Magnifica Humanitas, signed 2026-05-15 (the 135th anniversary of Rerum Novarum), promulgated 2026-05-25 — says so in 245 numbered sections: technology "is never neutral" (¶9), AI "must be disarmed" (¶110), lethal decisions may never be delegated to machines (¶198), and "the pursuit of greater profits cannot justify choices that systematically sacrifice jobs" (¶152). The v1 file, stamped 2026-05-26, said no such document existed.

Discipline & technical bet

Moral theology and canon law, delivered through the one instrument a pope uniquely has: binding teaching documents that 1.4 billion people (v1 figure, not re-verified) are asked to read. He makes no technical wager — the encyclical concedes "any statement regarding AI risks becoming quickly outdated" and that "all of us, including those who design them, possess only a limited understanding of their actual functioning"; systems are "more 'cultivated' than 'built'" (¶98). His anthropological wager is specific, though: AI "merely imitate[s] certain functions of human intelligence" and "this power remains entirely tied to data processing" (¶99) — a capability-skeptic premise the whole document rests on. The institutional bet is a new Curial organ: the Inter-Dicasterial Commission on Artificial Intelligence (ICAI), approved in audience 2026-05-03, established by rescript 2026-05-16, seven Vatican bodies, coordinated for its first year by Cardinal Czerny's Dicastery for Promoting Integral Human Development; first meeting 2026-06-17, now monthly (RNS, 2026-09-21). The only technical credential on the file is a 1977 mathematics degree — relevant, but not a bet.

Key claims (Says)

• AI is this papacy's central challenge (normative, signaling; 2025-05-10, address to the College of Cardinals): "I chose to take the name Leo XIV … mainly because Pope Leo XIII in his historic Encyclical Rerum Novarum addressed the social question in the context of the first great industrial revolution. In our own day, the Church offers to everyone the treasury of her social teaching in response to another industrial revolution and to developments in the field of artificial intelligence that pose new challenges for the defence of human dignity, justice and labour." Held — the falsifier v1 set ("if AI does not anchor major encyclicals … the framing thesis fails") was passed on 2026-05-25: the first encyclical is entirely about AI, and the first new Curial commission of the papacy is about AI (rescript 2026-05-16).
• Human dignity is the anchor concept (normative): Held — the encyclical's subtitle is "On Safeguarding the Human Person in the Time of Artificial Intelligence"; its stated criteria are "the inalienable dignity of the human person, the common good, the universal destination of goods, subsidiarity, solidarity and social justice" (¶96); "the human person is an end, not a means, and the economic order must remain subordinate to human dignity and the common good" (¶152). Restated to legislators 2026-08-21 and in the AI4Peace message (Vatican News, 2026-09-21).
• A "new Rerum Novarum" is needed (normative): Held — delivered. Signed on the anniversary; ¶3–4: "While Leo XIII spoke in his time of 'new things' (rerum novarum), today we cannot limit ourselves simply to repeating his insightful teachings." Chapter 4 ("Safeguarding Humanity at a Time of Transformation. Truth, Work, Freedom") is the labor chapter; the presentation speech calls AI "a transformation of similar magnitude, with perhaps even greater consequences."
• The Church has standing to speak globally (normative): Held as a claim; influence still unmeasured. Stated in his own terms 2026-05-25: "We do not possess technical answers, nor do we seek to displace those with expertise. But we bring a wisdom concerning the human that our present time desperately needs." Observable reception (see Track record): heads of government engaged (Carney phoned 05-29; Sánchez, Sheinbaum — Wikipedia's reception section, secondary), Bengio and Olah endorsed, the big labs said nothing (NBC, 2026-05-26), and Pedro Domingos called him "infallibly wrong."
• Continuity with Francis on AI ethics (empirical/normative): Held — the encyclical builds on Francis's technocratic-paradigm critique (¶92–93, Laudato Si'), quotes his G7 line that AI is "above all else a tool" (2025-06-17 message), and cites the Francis-era doctrinal note Antiqua et Nova (dated 14 January 2025) five times in its footnotes — a document v1 did not know about.
• AI's burdens will fall disproportionately on the poor (empirical, contested in degree): Held on Says, now with a two-branch mechanism. ¶153: "Wealthy societies automate rapidly and chaotically, reducing the need for a workforce … Vast regions of the world, by contrast, remain trapped in hybrid economies, where underpaid human labor and partial technologies coexist." ¶173 names the hidden labor chain — data labeling, content moderation, rare-earth extraction by children — and ¶178 calls data extraction a new colonialism. Empirical half: the rich-country branch has its first market-level support in the Canaries payroll data that regraded Brynjolfsson (see that file); the Global-South branch has no dataset attached yet.
• Subsidiarity applies to AI governance (normative): Held — v1's [TBD] resolved. ¶71: "The principle of subsidiarity applies especially in the context of the digital revolution. Here, the highest level is not the State, but rather major economic and technological actors that exercise de facto power over the conditions of everyday life." ¶72: "we cannot allow a handful of actors to dictate these processes on their own." ¶109: subsidiarity means "protecting the ability of communities to make choices and corrections, rather than confining their role to mere oversight after the standards have been set elsewhere."
• The "common good" supersedes shareholder logic (normative): Held — ¶152 (profits cannot justify sacrificing jobs); ¶151 on "outsized remuneration for a highly specialized minority alongside declining wages for a large portion of the workforce"; Crux interview (2025-07-30): "If we automate the whole world and only a few people have the means … there's a big problem, a huge problem coming down the line."
• NEW — "Technology is never neutral" (¶9): "technology is never neutral, because it takes on the characteristics of those who devise, finance, regulate and use it." Now the Holy See's standard line at the UN (Msgr Formica, 2026-06-19). Normative-descriptive hybrid; unfalsifiable as stated — noted, not graded.
• NEW — "Disarm AI" (¶110, and the presentation speech): "Disarming AI means freeing it from the mentality of 'armed' competition, which today is not limited simply to the military context, but is also an economic and cognitive phenomenon … To disarm means discrediting the assumption that technical power automatically confers the right to govern … merely regulating it is insufficient; it must be disarmed, welcoming and accessible." His own gloss on the word choice (2026-05-25): "The word is strong, I know, but deliberately chosen because this moment needs words capable of attracting attention." The vocabulary predates the encyclical — his World Day of Peace message (2025-12-08) is titled "Towards an 'unarmed and disarming' peace." A metaphor, not a mechanism; see Weak spots.
• NEW — Slow down, and politics must be able to (¶106): "Calling for prudence, rigorous evaluation and even, at times, a slower pace in adopting AI does not mean opposing progress … robust legal frameworks, independent oversight, informed users and a political system that does not abdicate its responsibility are required." Not a pause call — no "pause" or "moratorium" appears in the text — but America (2026-09-16) read it as anticipating the industry's own slow-down talk after a summer-2026 lab incident. The moratorium the Holy See does demand is narrower: lethal autonomous weapons (Conference on Disarmament, 2026-02-26).
• NEW — No machine may take a lethal or irreversible decision (¶198, ¶200): "it is not permissible to entrust lethal or otherwise irreversible decisions to artificial systems. No algorithm can make war morally acceptable." Non-negotiables: traceable decision chains, human control over lethal force, and "a shared framework — also at the international level — in order to curb the technological arms race." Preceded by the diplomatic-corps address (2026-01-09): "a danger of returning to the race of producing ever more sophisticated new weapons, also by means of artificial intelligence."
• NEW — Faces and voices are sacred; the challenge "is not technological, but anthropological" (World Communications Day message, 2026-01-24): "Behind this enormous invisible force that affects us all, there are only a handful of companies"; "Appropriate regulation can protect individuals from forming emotional attachments to chatbots"; LLM chatbots "are proving to be surprisingly effective at covert persuasion." Companion-AI regulation is now papal teaching, not just a Haidt / Harris position.
• NEW — Digital power now sits above the State (¶5, ¶95): "the main drivers of development are private, often transnational, parties that are endowed with resources and the capacity to intervene that surpass those of many Governments"; "control over platforms, infrastructure, data and computing power does not rest with States, but with major economic and technological actors." The load-bearing empirical premise under the whole governance argument; observable, and true on the 2026 evidence (labs out-spending most states on compute), though he offers no numbers.

Notable predictions — with falsifiable checks

• (2025-05-10) AI will be the defining technological and moral issue of his papacy. Held — first encyclical (2026-05-25), first new commission (2026-05-16), Communications Day message, Peace Day message, at least four papal messages to AI conferences, a UNESCO keynote booked. Check: by end-2027, is the second major teaching document (encyclical or apostolic exhortation) also AI-anchored, or does the papacy move on (war, migration, synodality)? A move-on is not a failure of the thesis, but it retires "defining."
• (2025-05-10, implicit) A major papal teaching document on AI will be issued. Held — fulfilled 2026-05-25 (Magnifica Humanitas, 42,300 words per Wikipedia; 245 sections and five chapters verified on the vatican.va text). v1's bracket "[TBD: confirm whether any draft has been signaled by 2026-05]" was already answered before the v1 review — OSV's chronology of 2026-05-14 calls the encyclical "coming," and the ICAI rescript (2026-05-16) was framed as preparation for it. Check retired; replacement check: does ICAI publish the Holy See's internal AI-use guidelines it was tasked with (humandevelopment.va, 2026-06-26) by 2027-06-30?
• (Inherited from Francis, 2024-06-14; restated ¶197–200, 2026-01-09, 2026-02-26) Autonomous lethal weapons require international restraint. Held as a position; outcome still open — no binding instrument; the CCW expert group closed three years of work on 2026-09-05 with 76 states supporting negotiations while "certain states, including Russia and the United States, weakened the text by insisting on changes to widely supported provisions" (HRW, 2026-09-07); the Holy See renewed its call for an immediate moratorium (Msgr Pacho, 2026-02-26; Abp Balestrero at the August session). Check: does the CCW Review Conference (November 2026) adopt a mandate to negotiate a legally binding instrument? Yes = the first hard win for this line; no = v1's "weak movement" stands for another cycle.
• (Implicit, ongoing) The Church becomes a more visible interlocutor in international AI governance. Held — upgraded from v1's "weak evidence". In-window: Holy See statement at the UN Independent International Scientific Panel on AI dialogue (2026-06-19); papal message read by the Permanent Observer at the ITU AI for Good summit (Geneva, 2026-07-08); a bipartisan US congressional delegation to the Vatican on AI ethics (Aug 2026, RNS 09-21); Abp Paglia scheduled for the Gates Foundation's Goalkeepers (2026-09-23); the first papal visit to UNESCO since 1980, with a keynote on AI (2026-09-25). Not found: any Holy See presence at the India AI Impact Summit (Feb 2026). Check: by 2027-06-30, does any government's AI law, strategy or summit declaration cite Magnifica Humanitas by name? (Canada's strategy landed the week after Carney's call — Wikipedia, secondary; the text itself was not checked.)
• (NEW, ¶151, 2026-05-15) "In some contexts, there is a legitimate fear of a significant and rapid contraction in available jobs." Framed as a fear, not a forecast — he leaves the number to others. Check: entry-level employment in AI-exposed occupations (the Canaries margin, −19% relative as of Aug 2026 per the Brynjolfsson file) — does the gap widen through end-2027? If it closes, ¶151's fear was misplaced and the encyclical's labor chapter over-weighted displacement.
• (NEW, presentation speech, 2026-05-25) AI "is also dramatically changing how war is waged." Empirical, stated in the present tense. Check: the CCW final report and any 2027 deployment audit of autonomous targeting; the sentence is Held if the November conference records deployed autonomous systems as a live issue rather than a prospective one.
• (NEW, ¶98) "any statement regarding AI risks becoming quickly outdated." A self-immunizing claim: it pre-excuses the document from capability forecasts, and the document makes none. Unfalsifiable — rewrite or retire; logged because it explains why the file has so few gradeable predictions.

Revealed behavior (Does)

• Spent the papacy's scarcest asset on AI. The first encyclical — the document a pope gets to write perhaps three or four times — went to AI, 12 months after election; John Paul II's first came in 1979, Francis's first in 2013, both on other subjects. Time and institutional capital, not adjectives.
• Presented it himself (Synod Hall, 2026-05-25) — popes normally delegate presentations to cardinals (Wikipedia; RNS 05-22 called it "unprecedented") — and put one AI-lab figure on the stage: "I would like to thank Mr Olah for accepting our invitation." Anthropic was the only frontier lab represented; the others "did not immediately respond" (NBC, 2026-05-26).
• Built an organ, not a committee: ICAI (rescript 2026-05-16), seven Curial bodies, one-year rotating coordination, tasked with mapping initiatives and writing the Holy See's own AI-use guidelines (humandevelopment.va, 2026-06-26); three meetings by September, monthly thereafter (RNS, 2026-09-21).
• Refused to be simulated. Asked to authorize an AI avatar of himself for "virtual audiences": "If there's anybody who should not be represented by an avatar, I would say the pope is high on the list … I'm not going to authorize that" (interview 2025-07-30, USCCB 2025-09-23). Says and Does match on the faces-and-voices line.
• Keeps a steady cadence of AI messages rather than one big speech: Rome AI/ethics/corporate-governance conference (2025-06-17); Builders AI Forum (2025-11-03 — notably pro-technology: "technological innovation can be a form of participation in the divine act of creation"); Peace Day (2025-12-08); diplomatic corps (2026-01-09); Communications Day (2026-01-24); Catholic legislators (2026-08-21: "Your task is not to hinder innovation but to guide it wisely"); AI for Good via Parolin (2026-07-08); AI4Peace (Vatican News, 2026-09-21).
• Travels where the labor chapter points: Monaco (2026-03-28), Algeria/Cameroon/Angola/Equatorial Guinea (2026-04-13 to 23), Spain (2026-06-06 to 12), France and UNESCO (2026-09-25 to 28), Uruguay/Argentina/Peru (2026-11-06 to 17) — vatican.va travels page. Africa before Paris.
• Names neither China nor the United States in an AI context — the encyclical's word "China" count is zero, and its two US mentions are the civil-rights movement (¶124) and the US bishops on work (¶167); "geopolitical or commercial dominance" (¶110) is as close as it gets. Consistent with the universalist claim; also consistent with a diplomat who needs both capitals. Meanwhile the first American pope is in an open rift with the US administration (Trump, 2026-04-12: "Pope Leo is WEAK on Crime") over Iran and Venezuela, not AI — Wikipedia, secondary.

Feels

"To disarm" is, in his words, "close to my heart" (¶110); the fear is dehumanization — "a paradox of material progress and anthropological regression" (¶154) — and the hope is the "civilization of love," a Paul VI phrase he uses as a program, not a slogan. Underneath: the 2017 Peru floods ("no one rebuilds alone," presentation speech) and a Nehemiah self-image — the administrator who rebuilds walls brick by brick, not the prophet who tears them down.

Hears

By his own account (2026-05-25): "scientists and engineers … political leaders and public officials … parents and teachers," plus "very troubling voices" on autonomous weapons and algorithmic exclusion. Institutionally: Czerny's dicastery and ICAI; Fr Paolo Benanti (the priest "advising two papacies" on AI, Fortune 2026-03-16); the Pontifical Academy for Life's Rome Call network; Francis's texts and Antiqua et Nova; the US bishops' AI task force (Bp Cantú). Intellectual diet in the encyclical: Augustine, Guardini (¶93), Tolkien (¶213), the whole social-encyclical line from Leo XIII to Francis. Downstream of no lab, no economist by name — the labor chapter cites bishops' conferences, not payroll data.

Sees

A reporting network no lab or state has — dioceses in every labor market the encyclical describes, from Chiclayo to Chicago — and the one vantage that made ¶173 possible: pastors who meet the data-labelers, moderators and miners. He also sees, uniquely, what heads of government say when they call the Vatican rather than the press. What he cannot see: inside a model, and he says so (¶98).

Incentive map

Re-derived this cycle. Commercial exposure: none — no equity, no fund, no vendor; CONVENTIONS rule 25 still does not apply, and this remains the bench's cleanest incentive map on money. Institutional incentives are strong and point one way: (1) relevance — a Church contested in the West gains standing by being the moral voice on the century's technology, so "AI is the defining issue" is also the papacy's best claim on attention (v1's framing-by-relevance caveat stands); (2) the Holy See is itself an AI adopter — one of ICAI's first tasks is internal-use guidelines — so the file should watch for "disarm" rhetoric softening as the Curia buys tools; (3) diplomatic neutrality — he must keep both Washington and Beijing, which is why the text names neither, and why "disarm" is aimed at a "mentality," not a country; RNS (2026-09-21) already notes Beijing may prove "more receptive" than Washington, which would make the Vatican's line look partisan whatever it intends. What he cannot say: that AI should be banned (¶4: technology "should not be considered, in itself, as a force antagonistic to humanity"; Builders AI Forum: innovation as "participation in the divine act of creation"); that any specific company is the villain (none is named anywhere); anything on enhancement that contradicts Church bioethics (¶115–116 handles transhumanism at the level of "narratives"); a capability forecast (¶98 forbids it). Political conflict: David Sacks — venture capitalist and the White House's AI adviser as of 2025 — answered the encyclical with "Who will guard the guardians?" (NBC, 2026-05-26); the Church–White House values gap is on record from the US bishops' own AI chair (Cantú, RNS 09-21). Prestige discount (rule 20): institutional weight gives reach; it validates none of the technical premises in ¶98–99.

Theories aligned with

• Adjacent to Automation Displacement — ¶150–154 is displacement-first: "current approaches to technology can paradoxically de-skill workers, subject them to automated surveillance and relegate them to rigid and repetitive tasks" (¶150); framed in dignity terms, not productivity-econ terms, as v1 said.
• Adjacent to Post-Labor Economics — ¶154 goes further than v1 expected: "we must rethink the nature of work and its connection to citizenship, ensuring that unemployment does not jeopardize social participation." Still no UBI (the words "basic income" do not appear); the remedy is "progressive redistribution and new rights of access to essential goods" via "intermediary communities" (¶153).
• Critical of Techno-Optimism — ¶115–116 names transhumanism and posthumanism as "the ideological background present in some centers of technological power"; ¶12: hope placed in "unlimited 'upgrades'"; ¶93: "More power does not necessarily imply something better."
• Adjacent to The Man Handed Over the Keys — the encyclical's version is ¶106: change "presented as necessary and inevitable, ultimately imposing rules shaped by those who control data, infrastructure and computing power," and ¶103 on handing an algorithm "the power to select who is worthy or not." Dignity framing, control-loss worry — the overlap v1 saw is now explicit text.

What he's reacting against

• Tech-industry framing of AI as morally neutral or inevitable — now the document's opening move (¶9 "never neutral"; ¶106 "presented as necessary and inevitable")
• Pure market governance of AI — ¶5: regulation is necessary but "the issue is not limited to regulation"; ¶152 on profits and jobs; the Leo XIII parallel is the whole point of the name
• US–China bipolar framing — reacted against by omission: neither country is named in an AI context; the enemy is a "mentality of 'armed' competition" (¶110)
• The Church being treated as irrelevant to frontier tech — "the Church wishes, with humility and frankness, to be part of conversations on artificial intelligence" (2026-05-25); the Olah invitation and the UNESCO stage are the tactic
• Pure doom and pure utopia — ¶14: "We cannot condone naïve enthusiasms, nor fuel unfounded fears"
• NEW: transhumanism/posthumanism as a worldview (¶115–116) — treated as a named ideological opponent "present in some centers of technological power," not as a symptom; the Vatican's on-record counter to Thiel's Antichrist lectures came from Benanti (essay 2026-03-14), not from Leo, who did not name Thiel (Fortune, 2026-03-16)
• NEW: attention-monetizing platforms and children (¶142) — "interventions by legislators are appropriate for setting age limits, holding service providers accountable rather than shifting the whole burden of control onto families"
• NEW: "artificial moral agents" — the idea that machines can judge right and wrong "with greater consistency than a human being" (¶198)

Where he overlaps / splits (with Rich)

• Overlaps with Shoshana Zuboff on extractive data capitalism — ¶178 "colonialism … appropriates data, transforming personal lives into exploitable information" is her thesis in papal voice; splits on register (sociological vs theological), as v1 said
• Overlaps with Yanis Varoufakis on capital concentration and labor dignity (¶67 lists "patents, algorithms, digital platforms, technological infrastructure and data" as goods "universally intended for everyone"); splits sharply on framework — Catholic social teaching is explicitly neither collectivist nor laissez-faire (¶31)
• Overlaps with Daron Acemoglu on institutional response over UBI and on the politics of deployment ("technological and organizational processes outside of democratic control," ¶154); splits on grounding — the encyclical cites no economist
• Overlaps with Kate Crawford more than v1 knew — ¶101 (AI systems "require enormous amounts of energy and water") and ¶173 (data labeling, content moderation, rare-earth mining by children) read as Atlas of AI summarized; splits on the source of moral authority
• Overlaps with Yoshua Bengio — now on record: "I agree with this sentiment by @Pontifex. The Vatican and other global institutions can and must play a role" (X, via NBC 2026-05-26); splits on primacy — Bengio leads with safety research, Leo with anthropology
• Overlaps with Chris Olah — shared a stage 2026-05-25; Olah asked for "informed critics who will tell the labs when we are failing"; splits on whether the answer is inside the model (interpretability) or outside it (law and conscience) — Leo says regulation itself is "insufficient" (¶110)
• Overlaps with Aza Raskin and Tristan Harris on "disarm" and on chatbots-as-persuasion (Communications Day 2026); Harris spoke at the Pontifical Urban University AI conference on 2026-05-21 (RNS 05-22); Raskin folded "disarm AI" into his own frame within days (per his bench file) — the three now amplify each other and are not independent confirmation
• Overlaps with Jonathan Haidt on age limits and platform accountability (¶142) — the encyclical is the largest institution yet to adopt that line; grade it against the regulator data in Haidt's file, not against papal endorsement
• Splits with Marc Andreessen on essentially all framing — now with an a16z reply on record: partner Eddy Lazzarin called the encyclical "defensive," protecting dignity "by saying intelligence was never the point" (NBC 05-26)
• Splits with Sam Altman on whose ethics govern deployment — Altman, Musk and Zuckerberg "did not immediately express public opinions" (NBC 05-26); silence is the split
• Splits with Peter Thiel — Thiel's Antichrist lectures cast one-world safety governance as the apocalyptic risk; ¶200 demands "a shared framework — also at the international level." The Vatican's reply came via Benanti, not Leo (Fortune 2026-03-16); the Thiel file over-credits a direct papal counter
• Adjacent to Tyler Cowen on institutional response time (¶106: "the frequent imbalance between the speed of technological growth and the slower development of awareness, norms, safeguards and institutions"); opposite valence on what institutions should optimize for
• Adjacent to Eric Schmidt only by proximity — a Minerva Dialogues participant with Vatican prelates (RNS 05-22); no papal endorsement or critique on record
• Overlaps with Frey / against Brynjolfsson on the labor sign — ¶151's "declining wages for a large portion of the workforce" and the "specialized minority" is a displacement-first reading; the Canaries data now sits behind it

Track record

• Pastoral track record in Peru re-dated: Chulucanas 1985–86 and Trujillo 1988–99 (Vatican News biography; Wikipedia says 1998), with a Chicago interval as vocation director between — Drifted from v1's continuous "1985–1998"; the substance stands: a Global-South bishop (Chiclayo 2014–2023) writing about "hybrid economies" of underpaid labor (¶153) has seen them
• Augustinian formation shows in the text (¶11 "our heart is restless," ¶130 the two loves, ¶215 justice and peace) — but v1's "[TBD: confirm if his Augustinian writings touch language/cognition]" found nothing: no published work on language or cognition surfaced this cycle; his only academic output located is the 1987 canon-law thesis. Per the Raskin rule, the credential leaves Track record and moves to Weak spots
• Prefect of the Dicastery for Bishops (2023-01-30 to 2025-05): high administrative experience, no public AI record before election — Held
• No publication record on AI prior to 2025-05 — Held; the record since is the thickest of any institutional persona on the bench: one encyclical, one commission, nine dated messages and addresses (2025-05-10 to 2026-09-21), one keynote pending
• Predecessor infrastructure inherited — Held and extended: Rome Call (2020), G7 (2024-06-14), Antiqua et Nova (2025-01), plus the International Theological Commission's Quo vadis, humanitas? (2026-02-09, cited in the encyclical's footnotes) — the Curia had two AI documents in the pipeline before the encyclical
• NEW: delivered the encyclical in 12.5 months, on the Rerum Novarum anniversary, and got Canada's prime minister on the phone within four days (Wikipedia, secondary) — as institutional execution, the fastest first encyclical with a live policy hook in the modern papacy; influence beyond phone calls remains unmeasured (see Weak spots)
• NEW: v1 credential error — "thesis at Catholic University of America [TBD: confirm topic — canon law]" — Wrong: the doctorate is from the Angelicum (1987), a plausible-sounding American guess that no primary supports
• Pattern: says exactly what the documents say, and then does it (avatar refusal, self-presentation, commission). The gap on this file is not Says-vs-Does; it is Says-vs-Effect.

Empirical vs normative

• Empirical: digital power now exceeds many states (¶5, ¶95); AI de-skills and surveils workers (¶150); a rapid job contraction is a "legitimate fear" (¶151); rich and poor economies get different harms (¶153); AI runs on hidden labor and resource extraction (¶173); autonomous weapons make war "more 'feasible'" (¶197); chatbots persuade covertly (2026-01-24). Mostly true on the 2026 evidence, none quantified by him
• Normative: dignity is the criterion; the person is an end; common good over profit; subsidiarity governs scale; lethal decisions stay human; the Church has standing; AI must be "disarmed"
• Anthropological premise (neither, and load-bearing): AI "merely imitate[s]" intelligence and is "entirely tied to data processing" (¶99). If frontier systems come to do things that look like judgment, the document's ethics survive but its description of the machine does not; no observable is attached and he has pre-excused the risk (¶98)
• Conflict note: no commercial exposure — rule 25 does not apply; institutional-relevance bias and rule-20 prestige apply, as in v1, now with one added conflict: the Holy See is itself procuring and governing AI use through ICAI

Weak spots / open questions

• v1: "Very early in papacy — almost all claims are framing-level; no encyclical-grade AI teaching document exists yet (as of 2026-05)." Wrong the day it was written — the encyclical was promulgated 2026-05-25 and the v1 review is stamped 2026-05-26; the document was publicly "coming" by 2026-05-14 (OSV). The May-2026 pass is confirmed defective on this file too
• v1: "Technical fluency on AI specifically is unclear … [TBD: track Vatican AI advisory roster]." Drifted — partly resolved, partly confirmed. Resolved: the roster is ICAI (seven Curial bodies, no outside technologists named on the rescript) plus Benanti; the encyclical concedes limited understanding (¶98). Confirmed: the technical critique landed — Domingos ("a series of ignorant and wrong-headed things about AI"; AI "will spread [power] more widely than ever"), Lazzarin ("light on the theology of artificial intelligence"), Sacks (NBC 05-26). Open: no external technical advisory body exists; Olah's "long collaboration" is a hope, not a structure
• v1: The Church's moral authority is contested in the West and the tech sphere — Held: eight companies declined comment (NBC 05-26); an evangelical leader denounced the encyclical's stance on capitalism and war (Moore; Wikipedia, secondary)
• v1: Real-world influence will take decades; near-term metrics are soft — Held, and now the metrics exist: three heads of government (Carney, Sánchez, Sheinbaum) and the US vice-president engaging with it (Wikipedia, secondary), one congressional delegation, one UNESCO stage, zero statutes. The check in Predictions (citation in any government AI text by 2027-06-30) is the file's attempt to make this hard
• v1: Catholic bioethics intersect awkwardly with enhancement/longevity/brain-interface debates — Drifted: the encyclical addresses transhumanism head-on (¶115–116) but at the level of "narratives"; the concrete cases v1 named are untouched
• v1: Augustinian vs Jesuit continuity not legible — Drifted — now legible: substance is Francis (technocratic paradigm, integral ecology, Antiqua et Nova), voice is Augustine (¶11, ¶130, ¶215); the first-person "I" and the Nehemiah/builder self-image are the difference, not the doctrine
• NEW: "Disarm" has no operational test. Freeing AI from "the mentality of 'armed' competition" (¶110) names no actor, threshold, treaty or metric; the only hard demand in the document is ¶200 (human control of lethal force, international framework). Until the UNESCO speech or ICAI attaches something checkable, the signature word is a mood
• NEW: the labor chapter's diagnosis (¶150–154, ¶173) is far stronger than its remedy ("progressive redistribution," "new rights of access," "intermediary communities") — the same gap Acemoglu is criticized for, without the economics
• NEW: no numbers anywhere in 42,300 words — no jobs figure, no energy figure, no date. This is deliberate (¶98) and it makes the encyclical nearly ungradeable; the file compensates with external observables
• NEW: US–China silence is a choice with a cost — RNS (2026-09-21) already frames Beijing as the more receptive capital; if the Vatican's line is adopted by one bloc and dismissed by the other, "universalist" becomes a partisan position by adoption (the Mazzucato rule: log the adopter)
• NEW: a secondary source (Wikipedia's article on the encyclical) says Leo declares just-war theory "outdated"; a full-text search of the fetched vatican.va English text finds no such sentence — "outdated" appears only about Paul VI's Octogesima Adveniens and AI statements (¶36, ¶98). Not carried; flagged so nobody imports it
• Carried from v1 as unresolved: Augustinian writings on language/cognition — none found; retire unless Rich wants a dedicated search of his Prior-General letters

Rich's take

• (your synthesis here)

Delta log

2026-09-21 — v1→v2 migration + validation (bulk run)

• Grades: 17 Held / 5 Drifted / 2 Wrong (plus 3 unfalsifiable items noted, not graded: ¶9 "never neutral" as stated, ¶98's self-immunizing clause, and the "disarm" metaphor as an operational claim; 1 v1 [TBD] retired unresolved — Augustinian writings on language).
• Wrong: v1 weak spot "no encyclical-grade AI teaching document exists yet (as of 2026-05)" — Magnifica Humanitas signed 2026-05-15, promulgated 2026-05-25 (vatican.va text; presentation address); v1 credential "thesis at Catholic University of America" — Angelicum, 1987 (Vatican News biography).
• Drifted: Peru dates (two stints, not 1985–1998 continuous — Vatican News biography); technical-fluency/advisory-roster weak spot (ICAI exists, rescript 2026-05-16; critique on record, NBC 2026-05-26); bioethics-intersection weak spot (¶115–116 addresses transhumanism, not the named cases); Augustinian-vs-Jesuit continuity (now legible in the text); v1's listing of the inaugural homily as an AI source — the 2025-05-18 homily contains no mention of AI, technology or work.
• Held: all eight v1 Key claims — papacy's central challenge (2025-05-10 address, encyclical); dignity anchor (¶96, ¶152); new Rerum Novarum (¶3–4, signed on the anniversary); standing to speak (presentation speech; reception per NBC and Wikipedia); Francis continuity (¶92–93; Antiqua et Nova footnotes); burdens on the poor (¶153, ¶173, ¶178); subsidiarity (¶71–72, ¶109 — v1 [TBD] resolved); common good over profit (¶151–152; Crux interview via USCCB, 2025-09-23). All four v1 predictions: defining issue; teaching document (fulfilled); LAWS restraint as position (HRW 2026-09-07; Crux 2026-02-26); governance-forum visibility (Holy See Geneva, 2026-07-08; Holy See UN statement, 2026-06-19; UNESCO, 2026-09-25; RNS 2026-09-21). Track record: Prefect role, no prior AI record, predecessor infrastructure. Weak spots: contested authority, soft influence metrics.
• Most surprising delta: the v1 "review" is dated one day after the encyclical it said did not exist — and the document's chosen verb is "disarm," which he says he picked "because this moment needs words capable of attracting attention" (2026-05-25). The bench learned about it from Aza Raskin's file before this one.
• Status/employer check (mandatory): alive, in office, publishing weekly through 2026-09-20; no health event found; travel per vatican.va travels 2026. New institution: ICAI (audience 2026-05-03, rescript 2026-05-16, first meeting 2026-06-17 per humandevelopment.va). Incentive map re-derived: still no commercial exposure; added the Holy See-as-AI-adopter conflict and the US-administration rift (Wikipedia, secondary).
• Signature-framing check: he still endorses the Leo XIII parallel (¶3–4, presentation speech) and added a new signature word ("disarm"); behavior matches (encyclical, commission, avatar refusal). No Says/Does divergence found.
• Unverified lead from the Raskin run — "amplifies a papal encyclical on AI" — confirmed against the primary: encyclical ¶110 and the presentation address both use "disarm"; the Raskin file's USCCB citation and 2026-05-25 date are correct.
• Tier: quarterly — a UNESCO keynote in four days, the CCW Review Conference in November, a Peru trip in November and monthly ICAI meetings make the next two quarters unusually dense; drop to semiannual once the November events are graded. Confidence raised from medium to medium-high: a 42,300-word primary now underwrites every Says line; influence claims remain the soft part.

Sources

• Primary — teaching documents: Encyclical Magnifica Humanitas (signed 2026-05-15, promulgated 2026-05-25) — paragraphs quoted: ¶3–5, 9, 12, 14–15, 67, 71–72, 92–93, 95–96, 98–99, 101, 103, 106, 109–110, 115–116, 130, 142, 150–154, 173, 178, 197–200, 213, 215, 229, 245 · Presentation and promulgation address, Synod Hall (2026-05-25) · Message for the 60th World Day of Social Communications, "Preserving Human Voices and Faces" (2026-01-24) · Message for the 59th World Day of Peace, "unarmed and disarming" peace (2025-12-08)
• Primary — addresses and messages: Address to the College of Cardinals (2025-05-10) · Inaugural homily (2025-05-18; no AI content) · Message to the Second Annual Rome Conference on AI, Ethics and Corporate Governance (2025-06-17) · Message to the Builders AI Forum (2025-11-03) · Address to the Diplomatic Corps (2026-01-09) · Address to the International Catholic Legislators Network (2026-08-21) · Leo XIV International Study Center (2026-09-18; no AI content) · Angelus (2026-09-20)
• Primary — institutions and diplomacy: Rescript establishing the Inter-Dicasterial Commission on AI (2026-05-16) · ICAI first meeting, 2026-06-17 (DPIHD, 2026-06-26) · Holy See statement, UN Independent International Scientific Panel on AI dialogue (2026-06-19) · Holy See at AI for Good, Geneva (2026-07-08) · Apostolic journeys 2026 · UNESCO — day of reflection on AI, education and peace (visit 2026-09-25)
• Biography: Vatican News biography (2025-05) · Wikipedia — Pope Leo XIV (citizenship 2015; Chiclayo to 2023; secondary)
• Reception and reporting (secondary): NBC — Tech titans mostly silent; Olah, Bengio, Dorsey, Sacks, Lazzarin, Domingos reactions (2026-05-26) · TIME (2026-05-25) · RNS — Silicon Valley turns to Rome (2026-05-22) · RNS — Vatican presses for global safeguards (2026-09-21) · America — what Leo has said about an "A.I. pause" (2026-09-16) · Vatican News — AI4Peace message (2026-09-21) · USCCB — Crux interview of 2025-07-30 (2025-09-23) · OSV — chronology of AI statements (2026-05-14) · Fortune — Thiel lectures and the Benanti response (2026-03-16) · Wikipedia — Magnifica humanitas (word count, chapter titles, government and church reactions; secondary) · Wikipedia — 2026 Holy See–United States rift (secondary)
• Autonomous weapons: HRW — CCW GGE concludes, 76 states for negotiations (2026-09-07) · Crux — Holy See renews moratorium call, Conference on Disarmament (2026-02-26) · Pope Francis, G7 Apulia address (2024-06-14)
• Catholic social-teaching foundation (carried from v1): Leo XIII, Rerum Novarum (1891-05-15); John Paul II, Laborem Exercens (1981); Benedict XVI, Caritas in Veritate (2009); Francis, Laudato Si' (2015), Fratelli Tutti (2020); Rome Call for AI Ethics (2020-02-28); Dicasteries for the Doctrine of the Faith and for Culture and Education, Antiqua et Nova (dated 2025-01-14 in the encyclical's footnotes); International Theological Commission, Quo vadis, humanitas? (2026-02-09, per footnote [123])
• v1 items resolved or retired: Vatican.va text of the 2025-05-10 address — linked above · inaugural homily — located, no AI content · Vatican biography — Vatican News, linked · Augustinian writings on language — not found, retired to Weak spots · CUA thesis — Wrong, corrected · tech-press reception — NBC, TIME, Fortune, RNS linked

migrated v1→v2 2026-09-21 · backup: _archives/pope-leo-xiv.html.bak-20260921 · converts-from: people/pope-leo-xiv.md · AI & Society domain