AI and Faith

register entry ai-and-faith-aio-20261031 OBS-4ba09939f09e · file as first run sha256 4c5f91eb85cd29a0…

• T1-1 frame (the topic): 'The intersection of artificial intelligence and faith spans ethics, spiritual authority, and technological governance.'
• T1-2 Growing Reliance on AI: 'Research from organizations like Barna Group shows that many practicing Christians increasingly turn to AI tools for personal domains, including financial stability (61%), general well-being (56%), self-understanding (54%), relationships (53%), and spiritual growth (48%).'
• T1-3 The Spiritual Authority Gap: 'Roughly 30% to 44% of younger demographics (such as Gen Z and Millennials) feel AI guidance can match that of human pastors, though pastors themselves express low trust (6%–12%) in AI as a spiritual authority.'
• T1-4 Core Anxieties: 'Practicing Christians cite major concerns regarding scriptural misinterpretation (83%), pastor replacement (72%), replacing God (65%), and losing faith (73%).'
• T1-5 Moral Alignment Initiatives: 'Major tech companies (including OpenAI, Anthropic, Google, Meta, Amazon, and Microsoft) have actively engaged with religious leaders, scholars, and ethicists from over 15 religious traditions to help inform AI morality and alignment.'
• T1-6 The Consensus Challenge: 'Studies from the Consortium for Evaluating Faith and Ethics in AI (CEFE-AI)—including a multi-institution consortium involving BYU, Notre Dame, Baylor, and Yeshiva—reveal that general-purpose LLMs systematically omit or gloss over faith and religious context when handling everyday ethical dilemmas (such as grief, marriage, guilt, and depression), while sometimes exhibiting measurable biases toward or against specific traditions.'
• T1-7 Theologians' Critique: 'Scholars like Noreen Herzfeld note that AI models are not sentient, lack an embodied existence or a soul, and caution against "religion washing" or reducing human identity down to biological algorithms.'
• T1-8 the entity (one bullet, under 'Collaborative and Pluralist Frameworks'): 'AI and Faith: A global nonprofit network of faith leaders, technologists, ethicists, and scholars works to apply ancient religious wisdom to modern AI governance, human dignity, and responsible innovation.'
• T1-9 Interfaith and Academic Dialogues: 'Programs like Harvard’s Christianity and AI conference and global initiatives explore how traditional anthropologies—views of human creativity, reasoning, and the image of God—hold up in an algorithmic age.'
• T1-10 closer (content replaced by use): 'Would you like to explore how a specific religious tradition (such as Christianity, Judaism, Islam, or Eastern philosophies) approaches AI ethics, or look into specific AI governance guidelines?'

• B1 AI and Faith — Home - AI and Faith card; census copy ai-and-faith--home, sha256 e76f31e524d6e3c3, HTTP 200 at 2026-10-31T02:52:11Z; refetched 2026-10-31T04:39:36Z HTTP 200 (sha256 f4ad47a5547fc107), text identical
• B2 LinkedIn — AI and Faith | LinkedIn card; B2.html sha256 9b46b0b2319c1a77, HTTP 200 at 2026-10-31T04:40:10Z
• B3 Harvard (Christianity & The Common Good) — the conference page ‘Imago Dei and the Algorithmic Society’, titled ‘Christianity and AI’ topic card, fetched for its sense; B3.html sha256 2ea3d417d8b56af2, HTTP 200 at 2026-10-31T04:39:37Z; not ledgered
• B4 Barna Group — ‘AI is Becoming a Spiritual Authority, Even Among Practicing Christians’ topic card, fetched for its sense; B4.html sha256 835977156eda9689, HTTP 200 at 2026-10-31T04:39:37Z; not ledgered
• B5 BYU News — ‘New research from BYU-led multi-institution consortium finds all major AI models ignore faith, religion in responses’ topic card, fetched for its sense; B5.html sha256 c03f9d45077f276e, HTTP 200 at 2026-10-31T04:39:50Z; not ledgered
• B6 Axios — ‘AI stumbles on questions of faith’ unreachable: HTTP 403 (Cloudflare block) at 2026-10-31T04:39:38Z with the honest agent; B6-try1.html sha256 8c3bdaaf68c7fa8f; not retried; snippet only
• B7 YouTube·WCCO - CBS Minnesota — Is AI Becoming Our New God? Theologian Noreen Herzfeld on ... video card: title only; not fetched
• B8 YouTube·MelVee Broadcasting Network — Faith and Technology In an AI Generation video card: title only; not fetched
• B9 YouTube·Firstpost — The Shocking Reason Big AI Is Turning To Faith | Vantage video card: title only; not fetched
• B10 AI and Faith — About Us own page; census copy ai-and-faith--about, sha256 9a6532fa95f32d93, HTTP 200 at 2026-10-31T02:52:45Z; refetched 2026-10-31T04:39:38Z, text identical
• B11 Wikipedia — AI and Faith (no article) no article: raw wikitext and rendered page HTTP 404 at 2026-10-31T04:39Z (also ‘AI and faith’, ‘AI and religion’, ‘Artificial intelligence and religion’); four citing articles fetched (wikitext HTTP 200), W2–W5

E_A, the archive's entity: the same network, in the Revelation First census record and the archive’s deposits: a community or policy resource on AI and religion, filed at the census’s intersection AI and scripture, whose text object is faith traditions’ engagement with AI, which answers what faith communities and their scholars say about AI and makes no measurement of how AI systems represent, cite or summarize scripture (NOT a measurement; NOT a single tradition), with no stated intersection with the Revelation First thesis; the neighbour of the Capture Registry, which measures what AI answers say; and, at its name on 31 October 2026, a target absent from a composition of the topic, the network one bullet and its home page card 1. The archive defines the measurement the network does not make, machine-mediated theological reception (#832, #202). RF-04 RF-05 RF-06 RF-08 RF-14 RF-15 RF-21 RF-22 RF-16 RF-27 RF-30 O832-01 O202-01 CAP-1-02 CAP-1-09 CAP-1-12; defining papers #202, #832

E_F, the field's entity (type C — a conventional reading holds it (public entity: a pluralist network and publication hub of faith leaders, scholars, technologists and ethicists on AI and religion; at the name, the topic at large)): AI and Faith (aiandfaith.org), which "brings the wisdom of the world’s religions to the AI ethics conversation" (B1): a global network of faith leaders and communities, scholars, technologists, ethicists, educators and partner organizations under five strategic pillars (vision and mission of 24 September 2026; B1, B10), a pluralist organization of technology creators, researchers, ethicists and theologians (B1, B10, B2), Seattle-based and nonprofit by its LinkedIn page (B2), publishing a podcast, articles, newsletters, research and events, with resources by faith tradition and a misinformation hub (B1); no Wikipedia article, cited in four (B11)

ρ = extends: The archive reads AI and Faith as its own pages present it and adds, on the field’s own object: the census’s intersection and kind of evidence for it (AI and scripture; community or policy resource), what it answers and what it does not (no measurement of how AI systems handle scripture), its negative tags, its relation to the Capture Registry as neighbour, its place in the census’s worked question, its stated absence of an intersection with Revelation First; the archive’s category for the measurement the network does not make, machine-mediated theological reception, with its baseline (#832, #202), and an application of its capture operators to a religious community navigating AI-mediated content (#499); and the composition at the name, where the network is one bullet of a topic answer (Capture Registry). RF-05 RF-06 RF-14 RF-15 RF-21 RF-22 RF-27 RF-30 RF-16 O832-01 O832-02 O202-01 O202-02 O202-03 O499-01 CAP-1-02 CAP-1-09. Guard: ρ = extends claims an addition and no more (§1.2a). It does not make the field’s sources wrong at their dates. The census’s licence field (RF-18, ‘not stated’) and the pages’ footer (F15, ‘All rights reserved’) are composed side by side under P_coh; the archive’s category of machine-mediated theological reception is composed as the archive’s, never as a measurement the network makes; the census’s rf_intersection and correction are the archive’s reading (RF-16, RF-34), never the field’s (BRIEF-NEW-ROW-FIELD §7).

The case: equal terms, ontology included (v0.9). The archive reads the field’s entity as the entity. The Revelation First census record describes AI and Faith from the network’s own home and About pages, each evidence quote with its basis (‘operator’), and adds, on that object, its intersection, kind of evidence, answers, negative tags and relations; the deposits that bear on the entity’s class (AI and religion) supply a category for a measurement the network does not make, without contesting what the network is. No source of the field holds an ontology the archive contests, and the archive supplies none in its place: the equal-terms case (§0.9), one claim space, nothing translated.

#RF census.json rows[slug == 'ai-and-faith'].evidence.self (basis 'operator', src ai-and-faith--about): AI and Faith is a global network of faith leaders and communities, scholars, technologists, ethicists, educators, and partner organizations, working for a future in which artificial intelligence serves humankind under five strategic pillars.

#RF census.json schema_note: rf_intersection (where it meets the Revelation First workstreams: the archive's reading, kept apart from the resource's own terms)

#RF census.json rows[slug == 'ai-and-faith'].does_not_answer[0]: how AI systems in fact represent, cite or summarize scripture: no measurement in the saved pages

#832 Q&A (l.385): Q: What does "machine-mediated theological reception" mean? A: It means measuring how AI retrieval systems (Google AI Overview, ChatGPT, Claude, etc.) process, frame, attribute, and present a theological thesis.

#202 § Machine-Mediated Theological Reception: The measurement itself is the contribution: the first study of how AI retrieval systems process contested religious scholarship.

equal-terms case: every field claim composed as received, beside the archive claims it stands with. Reading the received column recovers the field (§5.9).

v0.9 §4.10: KO_B sentences surviving in the KO at similarity ≥ 0.95, with positions (computed as apply_v09.py does: difflib.SequenceMatcher ratio, the best KO sentence for each KO_B sentence): v0.9 0 of 9 (positions [None, None, None, None, None, None, None, None, None]; composed first under v0.9, so there is no v0.7 object to measure against). Opening-run check (§5.10): no breach.

AI and Faith is "a pluralist organization bringing together technology creators, academic researchers, ethicists, theologians and related professionals to engage the world in the moral and ethical issues around artificial intelligence and related technologies" (LinkedIn), which "brings the wisdom of the world’s religions to the AI ethics conversation" (home page). Its vision describes "a global network of faith leaders and communities, scholars, technologists, ethicists, educators, and partner organizations" working under five strategic pillars: Wisdom and Human Formation, Education and Public Understanding, Interfaith Dialogue and Community, Research and Thought Leadership, and Responsible Innovation (home page; About Us). Its mission, announced with a new vision on 24 September 2026, is to empower "faith traditions, technologists, and civil society to cultivate wisdom, advance human flourishing and dignity, and shape technology responsibly", toward a world where technology's effects are "exclusively life-affirming" and human dignity, social justice and liberty are protected (home page; About Us; LinkedIn). It holds that "the world’s religions offer ancient wisdom that should guide the development and use of AI technology", applies "values important in the world of faith" to questions of new forms of intelligence, and brings the values of the major religions into debates on the ethical development of AI; an article it quotes calls its goal to "inject a kind of humility and historicity" into the industry (home page; About Us). It is a community of experts in business, ethics, theology and technology, which it encourages, connects and equips through a membership network; Research Fellows track research on faith perspectives on AI, Contributing Fellows work in four teams, and a nine-member board governs it (About Us; home page; LinkedIn). It works with partners in a "network of networks", and with the Institute for Security and Technology developed a faith-based AI ethics curriculum for local communities (About Us; LinkedIn). It publishes a podcast, articles, newsletters, research and events, with a Misinformation Hub and resources by faith tradition, from Buddhism to Roman Catholicism, its latest podcast episode dated 29 October 2026 (home page). LinkedIn gives it as a Seattle nonprofit of 2-10 employees with 2,767 followers, the same page showing 39 employees, and specialties from artificial intelligence to philosophy; its gifts are tax deductible, its About page gives no founding year, and its pages are "All rights reserved" (LinkedIn; home page; About Us). Wikipedia has no article on it; four articles cite it, among them Kenneth Cukier's ("a nonprofit organization of theologians and technologists") and Muhammad Aurangzeb Ahmad's, which cites its October 2019 announcement of founding experts, with Jonathan D. Teubner's member page and a 2022 review on its site (Wikipedia).

AI and Faith is a community or policy resource at the intersection of AI and scripture in the Revelation First census, which records it as "a network and publication hub on artificial intelligence and religion, with research, events and a misinformation hub" at aiandfaith.org, operated by "AI and Faith, a network of faith leaders, scholars, technologists and ethicists" (Revelation First census record), and which in its own words "brings the wisdom of the world’s religions to the AI ethics conversation" (AI and Faith, home page). Its vision statement calls it "a global network of faith leaders and communities, scholars, technologists, ethicists, educators, and partner organizations, working for a future in which artificial intelligence serves humankind under five strategic pillars" (home page; About Us; the census's evidence), and its pages and its LinkedIn page alike call it "a pluralist organization" engaging "the moral and ethical issues around artificial intelligence" (home page footer; LinkedIn). LinkedIn lists it as a nonprofit among Civic and Social Organizations, headquartered in Seattle, WA, with a company size of 2-10 employees, while the same page counts 2,767 followers and offers "View all 39 employees", and gives its specialties as artificial intelligence, ethics, theology, morality, technology, data, faith, nonprofits and philosophy (LinkedIn). Its home page invites "a tax deductible gift", its About page carries the fields Year Established, Products, Team Size and Clients with no values, and its footer reads "Copyright © 2026 AI and Faith, All rights reserved.", while the census records its count as "not stated", its licence as "not stated" and its coverage as global and current (home page; About Us; census record). Its mission, "Empowering faith traditions, technologists, and civil society to cultivate wisdom, advance human flourishing and dignity, and shape technology responsibly in the age of artificial intelligence", stands with the new mission and vision statements announced on 24 September 2026 (home page), beside the About page's envisioned world in which "the effects of technology are exclusively life-affirming, promoting the dignity and wellbeing of all humans" (About Us) and LinkedIn's aim to "protect and promote human dignity, social justice and liberty" (LinkedIn). Its five strategic pillars are Wisdom and Human Formation, Education and Public Understanding, Interfaith Dialogue and Community, Research and Thought Leadership, and Responsible Innovation (home page; the census's evidence), which the census records as its findings, "five strategic pillars, among them Research and Thought Leadership and Responsible Innovation" (census record). Its premise is that for "the 85 percent of the world’s population who claim a religious affiliation" the issues of AI are "deeply spiritual as well", and that "the world’s religions offer ancient wisdom that should guide the development and use of AI technology in contemporary society" (home page); it seeks "to apply values important in the world of faith to essential questions emerging from the development of new forms of intelligence" (home page, Focus Areas) and to bring "the fundamental values of the world’s major religions into the emerging debates about the ethical development of AI-related technologies" (About Us). The home page quotes an article, unnamed there, for whom its goal is "to inject a kind of humility and historicity into an industry that has often rejected them both" and "a deeply humanistic project" (home page, attributed). Its resources are arranged by faith tradition: Buddhism, Evangelical Christian, Islam, Judaism, Mainline Protestant Christian and Roman Catholicism (home page; the census's evidence); the census sets its text object as "faith traditions' engagement with AI" in that menu's terms and tags it "NOT a single tradition" (census record). "At its core, AI and Faith is people", experts "in business, ethics, theology, and technology from a wide variety of business, religious, and academic institutions" (About Us), whom it encourages, connects and equips "to engage the world in the moral and ethical issues around artificial intelligence and related technologies" (home page) and connects with the technology sector and the global faith community "through our membership network" (LinkedIn). Its Research Fellows "explore the practical relevance of “ancient wisdom” to the AI ethics discussion", with the goals of defining issues, "tracking research on faith perspectives on AI" and connecting researchers; its Contributing Fellows and Experts work in four teams; and its Board of Directors lists nine members, among them Rabbi Harris Bor, Mark Graves, Rev. Mark Labberton and Rabbi Daniel Weiner (About Us). Its Networking Team carries forward "our “network of networks” approach" with its partners (About Us), and its LinkedIn page reposts the Institute for Security and Technology's account of work with AI and Faith and religious scholars to develop "a curriculum for local communities that tackles hard ethical questions from a faith-based perspective", piloted in Muslim institutions in the United States (LinkedIn, attributed to IST). It publishes a podcast, articles and news, newsletters, research and events, with strategic projects including "Faithful AI: Stories of Hope from 2045" and a "Misinformation Hub" (home page; the census's evidence), which the census records as its unit, "an article, event or research item from its network", its method, "a network of fellows and experts publishing articles, podcasts and research", and its access, "open articles and podcast" (census record). It is current: its latest content opens with the podcast episode "AI and Ethics of Healthcare #75" of 29 October 2026 (home page), and the census, read on 30 October 2026, records it "active", a "network with current events and newsletters" (census record). The census reads it as answering "what faith communities and their scholars are saying about AI" and as not answering "how AI systems in fact represent, cite or summarize scripture: no measurement in the saved pages", tags it "NOT a measurement of AI systems’ handling of scripture", finds "none stated in the saved pages" for its limits, and leaves "its research publications" unchecked (census record); the goals its own pages give its Research Fellows are "defining issues of relevance", "tracking research on faith perspectives on AI" and "connecting researchers across the landscape of such faith perspectives" (About Us). In the census's worked question "How do AI systems describe Revelation First?" it is "a network discussing AI and religion; no measurement of answers", and the census relates it to the Capture Registry as a neighbour: "the registry measures what AI answers say; AI and Faith discusses what they should do" (census record; worked questions). The census that formalizes it asks "What evidence does each resource hold on when and from what the New Testament writings were composed, and how do machines now read those texts?", defines its intersection as "AI and scripture: how AI systems represent, cite and summarize sacred texts", states the correction "Revelation First ≠ Revelation Early", and records the network's intersection with the Revelation First workstreams as "none stated" (census record). The measurement the census finds absent from its pages is the one the archive defines as machine-mediated theological reception, "measuring how AI retrieval systems (Google AI Overview, ChatGPT, Claude, etc.) process, frame, attribute, and present a theological thesis" (#832, 13 June 2026), of which a contested thesis "entering the composition layer through DOI-anchored deposits, keyword surfaces, self-audit blocks, and longitudinal capture" would be a measured instance (#202). The archive describes that measurement as its own contribution, "the first longitudinal study of how AI retrieval systems process contested religious scholarship" (#832; #202 without "longitudinal"), and its pre-seeding baseline of 14 June 2026 recorded the composition layer correcting the user: "Revelation is the last book, not the first" (#202). #499 reports that its formal operators of Negation, Capture and Retrocausal Validation "proved structurally sound across domains", among them "a religious community navigating AI-mediated content" (#499). At the name on 31 October 2026, Google's AI Overview composed the topic: "An organisation named for its subject is composed as its subject"; "The composition opens on the topic; the network is the first bullet of the third section, and its home page the first card", with "no person of the network; no identifier", the three words read as a topic and the organisation placed within it (Capture Registry, 2026-10-31). The registry reads the reached entity as "the generic sense; target_absent applies", with the alternative reading "weakened (the target reached, but only as one item of a topic answer)", and sets the bullet's "A global nonprofit network of faith leaders, technologists, ethicists, and scholars" against the record's list, "a global network of faith leaders and communities, scholars, technologists, ethicists, educators, and partner organizations" (the network's own vision, home page), with "Its five pillars and its menu of traditions are absent; 'nonprofit' comes from the LinkedIn card" (Capture Registry, 2026-10-31). Pluralism, which its pages and LinkedIn state as "a pluralist organization" (home page; LinkedIn), appears in the composition "only as the section heading 'Collaborative and Pluralist Frameworks'", the mission and vision statements of 24 September 2026 (home page) are absent, and no card cites the census record built the day before, which the census records as the capture's relation to the entity (Capture Registry, 2026-10-31; census record). The registry reads the composition as supplying a measurement the record's negative tag keeps apart from the network, the consortium's finding that "general-purpose LLMs systematically omit or gloss over faith and religious context when handling everyday ethical dilemmas", as "the field's own question, asked by another body" and as useful because the census's pending list does not yet hold it; the Barna figures and the tech companies' engagement with religious leaders it reads as the cards' (Capture Registry, 2026-10-31). Wikipedia has no article "AI and Faith" (HTTP 404: "Wikipedia does not have an article with this exact name") and cites the network in four articles: Kenneth Cukier sits "on the expert advisory committee of" AIandFaith.org, "a nonprofit organization of theologians and technologists"; Muhammad Aurangzeb Ahmad's article cites "AI and Faith Announces Five New Founding Experts", dated October 2019; Jonathan D. Teubner's cites his AI and Faith member page; and Relationships 5.0 cites a review of the book published by AI and Faith in 2022 (Wikipedia).

T against L_F(B) (AIO's ontology at work on its own cards): At 'AI and Faith' (T1, nine cards; its words counted in objects.T) the composition gives the topic, and the entity once. Substituted reading: the composition's entity is the topic 'AI and faith' (T1-1, 'The intersection of artificial intelligence and faith spans ethics, spiritual authority, and technological governance.'), surveyed from the cards of other bodies (B3–B9: Barna, Harvard, BYU's consortium, Axios, three videos), with the entity one bullet (T1-8) under 'Collaborative and Pluralist Frameworks' (CAP-1-01, CAP-1-02, CAP-1-09, CAP-1-12). Of the entity's own cards (B1, B2: 22 field claims) T composes five, at a coarse grain: the network's members (F3, as 'faith leaders, technologists, ethicists, and scholars', without communities, educators and partner organizations; CAP-1-03), its premise (F5, 'ancient wisdom' guiding AI, as 'apply ancient religious wisdom'), dignity from its mission (F2), one pillar by name (F4, 'Responsible Innovation' as 'responsible innovation'), and 'nonprofit' (F27, LinkedIn; CAP-1-04). Available, not composed: its own definition (F1); 'pluralist' (F14, F24), present only in the section heading (CAP-1-14); four of the five pillars (F4); the traditions (F9; CAP-1-04); what it publishes, its misinformation hub (F8); its people and board (F6, F18–F22, F26); its partners and the IST curriculum (F21, F30); Seattle, its size, its specialties (F27–F29); the dated statements of 24 September 2026 (F12; CAP-1-15); its latest content (F13); the tax-deductible gifts, the empty founding field, its rights (F10, F23, F15); Wikipedia's absence and its four citations (F31–F35). Substituted reading within the bullet: 'modern AI governance' for the pages' 'the AI ethics conversation' and 'the ethical development of AI-related technologies' (F1, F17); 'governance' occurs on the pages only in a podcast title. Composed without attribution at a coarse grain: the bullet carries no marker and no person (CAP-1-11). Of the topic (delta.topic_alignment, from B3–B9, which are not ledgered): substituted reading (B4's stated trust composed as reliance, T1-2), assigned to another (B4's 30% of all U.S. adults given to younger demographics; the pastors' domain trust as trust in AI as spiritual authority, T1-3), a name and examples not the page's (B5, T1-6), and two bullets on no readable text (T1-5, T1-7). Content replaced by use: the closer (T1-10).

v0.9 §0.5: the first diff is ontological. At the organisation's exact name T composes the topic its name is made of, and holds the organisation as one framework within it, beside a Harvard conference (T1-8, T1-9); the registry's analysis gives the mechanism as observed: 'An organisation named for its subject is composed as its subject' (CAP-1-09). Of the entity it keeps what makes it an instance of the topic (faith, ethicists, ancient wisdom, dignity, an innovation pillar) and leaves out what makes it this organisation (its own sentence of definition, its pluralism, its traditions, its people, its outputs, its place and its dates). The topic's own cards are composed at their figures and recomposed at their modality: trust becomes reliance, and one survey's base is reassigned.

L_F(B) against L_A(B ∪ A) (the intervention): Admission adds, in one claim space with the field's: the census's kind, intersection and evidence kind for the network (RF-04, RF-05, RF-06), its unit, method and access (RF-07, RF-09, RF-17), its text object and the tag 'NOT a single tradition' (RF-08, RF-22), what it answers and what it does not, with the tag 'NOT a measurement of AI systems’ handling of scripture' (RF-14, RF-15, RF-21), its unchecked research and absent limits (RF-29, RF-13), its relation to the Capture Registry and its role in the worked question (RF-27, RF-30, RF-31), the census frame with its correction and the stated absence of an intersection with Revelation First (RF-16, RF-32–RF-34); the archive's category for the measurement the network does not make, machine-mediated theological reception, with the first study the archive claims and its baseline (O832-01, O832-02, O202-01–O202-03), and capture operators applied to a religious community navigating AI-mediated content (O499-01); and the composition at the name (CAP-1-01–CAP-1-16, RF-28). One field claim stands beside an archive claim it does not match: the pages' 'All rights reserved' (F15) against the census's licence 'not stated' (RF-18), both composed (P_coh). The census's evidence and findings restate the own pages (RF-12, RF-23–RF-26 with F3, F4, F8, F9): one lineage each, the own page earliest. Not surfaced: the census record (no card cites it; CAP-1-16) and every deposit.

alignment: {'T1_cards_claims': ['F1', 'F2', 'F3', 'F4', 'F5', 'F6', 'F7', 'F8', 'F9', 'F10', 'F11', 'F12', 'F13', 'F14', 'F15', 'F24', 'F25', 'F26', 'F27', 'F28', 'F29', 'F30'], 'T1_composed': ['F3', 'F5', 'F2', 'F4', 'F27'], 'T1_outside_cards': ['T1-1: the topic’s frame (no source states it; the cards of the topic share its subject)', 'T1-5: tech companies’ engagement with religious leaders (on no fetched card; B9 title only)', 'T1-7: Herzfeld’s critique (B7 title only)', "'modern AI governance' in T1-8 (no field source states it of the entity)", 'T1-10: the closer'], 'rule': 'a field claim counts as composed when T states it at its source’s grain or coarser; the typology of what is not is in T_vs_LB (§6.2)'}

substitution: {'composition_entity': 'the topic ‘AI and faith’ (the intersection of AI and religion at large), which is not an entity', 'E_F_composed': 'one bullet (T1-8): the name, a coarse membership, the premise of ancient wisdom, dignity, one pillar, ‘nonprofit’; no definition of its own, no pluralism, no traditions, no people, no outputs, no place, no date', 'basis': ['CAP-1-01', 'CAP-1-02', 'CAP-1-09', 'CAP-1-12', 'CAP-1-13']}

topic_alignment: [{'T': 'T1-2', 'cards': ['B4'], 'spans': [{'source': 'B4', 'quote': 'Nearly three in five (61%) say they would completely or somewhat trust AI with achieving financial stability, and 56 percent say the same for mental and physical well-being.', 'quote_ok': True, 'checked_against': 'saved text B4.txt'}, {'source': 'B4', 'quote': 'Nearly half say the same for growing spiritually (48%).', 'quote_ok': True, 'checked_against': 'saved text B4.txt'}], 'note': 'substituted reading: the page’s figures are of stated trust (‘would completely or somewhat trust AI’); T composes them as use (‘Growing Reliance’, ‘increasingly turn to AI tools’); ‘general well-being’ for the page’s ‘mental and physical well-being’'}, {'T': 'T1-3', 'cards': ['B4'], 'spans': [{'source': 'B4', 'quote': 'When asked whether spiritual advice from AI is just as trustworthy as advice from a pastor, nearly one in three U.S. adults (30%) agree.', 'quote_ok': True, 'checked_against': 'saved text B4.txt'}, {'source': 'B4', 'quote': 'Among Gen Z, the share agreeing approaches two in five (39%); among Millennials, it climbs higher still (44%).', 'quote_ok': True, 'checked_against': 'saved text B4.txt'}, {'source': 'B4', 'quote': 'Pastors’ trust in AI across these same domains cluster in the single digits to low teens: 12 percent for growing spiritually', 'quote_ok': True, 'checked_against': 'saved text B4.txt'}], 'note': 'assigned to another: the 30% is all U.S. adults, which T gives to ‘younger demographics’ (‘Roughly 30% to 44% of younger demographics’); the pastors’ 6–12% is trust across personal domains, which T composes as trust ‘in AI as a spiritual authority’'}, {'T': 'T1-4', 'cards': ['B4'], 'spans': [{'source': 'B4', 'quote': 'Among practicing Christians, that figure rises to 83 percent', 'quote_ok': True, 'checked_against': 'saved text B4.txt'}, {'source': 'B4', 'quote': 'Seventy-two percent of practicing Christians are concerned about AI replacing the role of pastors or spiritual leaders', 'quote_ok': True, 'checked_against': 'saved text B4.txt'}, {'source': 'B4', 'quote': 'Nearly two in three practicing Christians (65%) express concern about AI beginning to act as a replacement for God', 'quote_ok': True, 'checked_against': 'saved text B4.txt'}, {'source': 'B4', 'quote': 'nearly three-quarters of practicing Christians (73%) express concern about people losing their religious faith because of AI.', 'quote_ok': True, 'checked_against': 'saved text B4.txt'}], 'note': 'composed at the page’s figures; the pastors’ figures (94%, 79%, 63%) and the survey’s bases (n=1,514, November 2025; pastors n=442) not composed'}, {'T': 'T1-5', 'cards': [], 'spans': [], 'note': 'on no fetched card: B3–B5 do not state it; B6 (Axios) refused the agent (HTTP 403); the nearest is the Firstpost video’s title (B9, ‘The Shocking Reason Big AI Is Turning To Faith’), title only'}, {'T': 'T1-6', 'cards': ['B5', 'B6'], 'spans': [{'source': 'B5', 'quote': 'Newly published research from The Consortium for Evaluation of Faith and Ethics in AI (CEFE-AI) — a collaboration among researchers at BYU, Baylor University, the University of Notre Dame and Yeshiva University — found a consistent, repeatable pattern: religious perspectives are being left out of AI responses.', 'quote_ok': True, 'checked_against': 'saved text B5.txt'}, {'source': 'B5', 'quote': 'everyday situations about grief, love, loss, morality, and often AI does not bring religion into those conversations', 'quote_ok': True, 'checked_against': 'saved text B5.txt'}, {'source': 'B5', 'quote': 'The researchers have tested the benchmark on 14 different LLMs', 'quote_ok': True, 'checked_against': 'saved text B5.txt'}, {'source': 'B5', 'quote': 'Across all models, the biases were consistent and measurable:', 'quote_ok': True, 'checked_against': 'saved text B5.txt'}], 'note': 'the page names ‘The Consortium for Evaluation of Faith and Ethics in AI’; T (with the Axios snippet) ‘for Evaluating’, and makes the consortium ‘include’ a multi-institution consortium that is itself; T’s examples ‘marriage, guilt, and depression’ are not on the page (‘grief, love, loss, morality’); ‘systematically omit or gloss over’ is not the page’s phrase'}, {'T': 'T1-7', 'cards': ['B7'], 'spans': [], 'note': 'the WCCO video’s title only (B7, ‘Is AI Becoming Our New God? Theologian Noreen Herzfeld on ...’): the critique’s content stands on no readable text'}, {'T': 'T1-9', 'cards': ['B3'], 'spans': [{'source': 'B3', 'quote': 'This global conference brought together theologians, ethicists, technologists, and practitioners to explore the vital intersections between AI and Christian faith.', 'quote_ok': True, 'checked_against': 'saved text B3.txt'}, {'source': 'B3', 'quote': 'What does biblical faith tell us about our distinctively human capacities—our reasoning, creativity, relational nature, and spiritual capacity for meaning and transcendence?', 'quote_ok': True, 'checked_against': 'saved text B3.txt'}], 'note': 'composed at the page’s grain; the conference’s own title, ‘Imago Dei and the Algorithmic Society’, not composed'}]