The Information Machine

Anthropic's Push to Broaden AI Values Input · history

Version 8

2026-05-25 20:44 UTC · 154 items

What

Pope Leo XIV's first papal encyclical, 'Magnifica Humanitas,' has been published — a 42,300-word document calling for robust AI regulation and elevating AI ethics from corporate self-governance to a religious imperative [21][24][22]. The encyclical arrives as Anthropic remains banned from U.S. federal use after refusing Pentagon demands to disable its AI safeguards [1][6], having lost the resulting government contract to OpenAI [8][9], and seeing its published moral framework — Claude's constitution — dismissed by an Oversight Board member as 'about vibes, not rights' [15]. Both Anthropic and OpenAI are consulting the same religious and cross-cultural communities for AI values input [10][17][18], a convergent industry practice the Vatican's call for binding regulation implicitly frames as structurally insufficient [23].

Why it matters

The encyclical closes the loop on a months-long escalation: Anthropic built a voluntary ethics framework, consulted religious communities, and published Claude's constitution — and now the most authoritative religious institution in the world has responded by calling for binding government regulation rather than corporate codes. The story's fault line has shifted from whether Anthropic's approach is sincere to whether any voluntary ethical framework, however thoughtfully constructed, can substitute for enforceable law — a question that now carries the weight of formal Catholic doctrine.

Open questions

  • What specific regulatory mechanisms does 'Magnifica Humanitas' propose, and does it name binding obligations for governments vs. AI companies directly? [21][24][22]

  • Will the encyclical's call for robust regulation accelerate legislative action in the EU, U.S., or other jurisdictions — and does it strengthen or complicate Anthropic's position as a values-based actor still facing a federal ban? [24][6]

  • Will Anthropic revise Claude's constitution in response to the Oversight Board's 'vibes not rights' critique [15] and the Vatican's published demand for binding regulation rather than ethical codes [23]?

  • With both Anthropic and OpenAI consulting the same religious communities [17][18] and the Vatican now calling for regulation rather than codes [24], does AI religious consultation remain a credible ethics differentiation strategy — or has it become an industry legitimacy credential?

Narrative

Anthropic's conflict with the U.S. government began when the Pentagon demanded the company disable its AI safeguards as a condition for military use. The company's CEO explicitly refused [1], and the Trump administration responded by banning Anthropic from federal use [2] and designating it a supply chain risk. Catholic moral theologians filed an amicus brief in case 26-1049 backing the refusal on human dignity grounds [3], and a federal judge temporarily blocked the Pentagon designation [4][5] — but a federal appeals court overturned that reprieve on April 8, 2026, leaving the ban in effect [6][7]. OpenAI moved within hours of the Trump ban announcement to claim the Pentagon AI contract [8][9], producing a direct market demonstration that values-based refusals carry quantifiable competitive costs when a less restrictive competitor is available.

Against this legal backdrop, Anthropic pursued a formal values framework for its AI systems. In May 2026, the company disclosed consultations with more than fifteen religious and cross-cultural traditions [10] and released Claude's constitution — described by Vox as approximately 80 pages [11], with architect Amanda Askell profiled by the WSJ, Vox, and Der Spiegel [12][11][13]. The institutional response was divided: Oxford AI Ethics offered academic analysis [14], while Oversight Board member Suzanne N. dismissed the document as 'a constitution that is about vibes, not rights' [15][16] — a direct challenge to whether it carries enforceable weight. Multiple sources confirmed that OpenAI is independently consulting Hindu, Sikh, and Christian leaders in parallel [17][18], situating Anthropic's consultation program within an industry-wide pattern rather than a distinctive strategy [19].

The Vatican's response escalated from informal warnings to formal magisterial teaching. Pope Leo XIV had previously warned that children and adolescents are vulnerable to AI manipulation [20]; on May 25, 2026, his first papal encyclical — titled 'Magnifica Humanitas' and running to 42,300 words — was published [21][22]. The Washington Post framed it as elevating AI ethics to a religious imperative [23]; PBS reported its central demand as a call for robust AI regulation [24]. The document represents the most authoritative institutional argument yet that voluntary ethical codes are structurally insufficient and that binding regulatory frameworks, not corporate constitutions, are required. The Rome Call for AI Ethics, a Vatican-led interfaith document with Muslim, Jewish, and evangelical signatories [25], provides the broader formal framework within which the encyclical sits. Whether 'Magnifica Humanitas' reshapes public policy debates or alters the position of AI companies that have invested in ethics consultation strategies is now the central unresolved question.

Timeline

  • 2025-11: Pope Leo XIV warns that children and adolescents are vulnerable to AI manipulation and states the real danger of AI is not technology itself [20][41]
  • 2026-02-27: Trump administration bans Anthropic from Pentagon use; OpenAI announces a Pentagon AI deal to fill the resulting contract gap within hours [8][9][2]
  • 2026-03-16: Catholic moral theologians file amicus brief (case 26-1049) in the U.S. Court of Appeals backing Anthropic's refusal to comply with Pentagon demands [3]
  • 2026-03-26: Federal judge temporarily blocks the Pentagon from branding Anthropic a 'supply chain risk' and halts the Trump administration's ban on federal use [4][5]
  • 2026-04-08: Federal appeals court overturns the temporary block, declining to prevent the Pentagon's blacklisting of Anthropic [6][7]
  • 2026-04-11: Washington Post reports Anthropic consulted Christian leaders for advice on Claude's moral future [42]
  • 2026-04-20: New York Times publishes opinion piece questioning whether religion is the right answer to AI morality [35]
  • 2026-05-19: Anthropic publishes 'Widening the conversation on frontier AI,' describing dialogues with 15+ religious and cross-cultural traditions [10]
  • 2026-05-21: Skeptical 'beautiful PR' framing spreads across social media; Hacker News thread on the initiative opens [33][34][43]
  • 2026-05-23: WSJ, Vox, and Der Spiegel profile Amanda Askell as Claude's moral framework architect; alignment-faking research and Alignment Forum critique circulate [12][11][13][38][40]
  • 2026-05-24: Anthropic publishes Claude's constitution publicly; Oversight Board member calls it 'about vibes, not rights'; multiple sources confirm OpenAI is also consulting religious leaders in parallel [27][15][16][17][18]
  • 2026-05-24: BBC reports Anthropic CEO explicitly rejected Pentagon demands; Seattle Times frames tech's religious consultations as an industry-wide pattern; Nicholas Thompson calls Claude's constitution 'the most interesting thing in tech' [1][19][44]
  • 2026-05-25: Pope Leo XIV's first papal encyclical 'Magnifica Humanitas' (42,300 words) is published, calling for robust AI regulation and elevating AI ethics to a religious imperative [23][21][24][22]

Perspectives

Anthropic

Maintained a values-based refusal of Pentagon demands to drop AI safeguards [1]; lost a federal appeals court challenge [6][7] with OpenAI filling the contract gap [8]; published Claude's constitution, dismissed by the Oversight Board as 'about vibes, not rights' [15] and now challenged by the Vatican's call for binding regulation [23]

Evolution: Compounding institutional pressure: the adverse court ruling, OpenAI's contract capture, the Oversight Board critique, and the Vatican's 'Magnifica Humanitas' all converge on the same structural challenge to Anthropic's code-based ethics approach — with no public Anthropic response to any of these institutional challenges

[1][26][15][27][8][6][7][23]

Trump administration / U.S. Department of Defense

Banned Anthropic from federal use after the company refused to disable AI safeguards [2]; branded the company a supply chain risk; prevailed in federal appeals court when Anthropic challenged the blacklisting [6][7]

Evolution: Consistent; ban stands and the DoD has a replacement vendor in OpenAI [8]

[8][2][6][7][1]

OpenAI

Claimed the Pentagon AI contract within hours of the Trump ban on Anthropic [8][9]; also independently consulting Hindu, Sikh, and Christian religious leaders for ethical guidance [17][18], positioning OpenAI as both the direct market beneficiary of Anthropic's refusal and a parallel actor in religious consultation

Evolution: Consistent; the convergence of both companies on religious consultation strengthens the 'industry credential' interpretation of the practice

[8][9][17][18]

Pope Leo XIV / Vatican

Published 'Magnifica Humanitas,' a 42,300-word encyclical calling for robust AI regulation and elevating AI ethics to a religious imperative [21][24][22]; framed by the Washington Post as the strongest institutional challenge yet to corporate AI self-governance [23]

Evolution: Decisively escalated: from informal papal warnings [20] and Papal Academy previews [28] to formal magisterial teaching — the Vatican's most authoritative intervention — now on record demanding binding regulation rather than ethical codes

[23][21][24][22][20][29][28]

Oversight Board

Member Suzanne N. characterized Claude's constitution as 'a constitution that is about vibes, not rights' [15][16], issuing a direct indictment of the document's enforceability rather than a call for iterative improvement

Evolution: Consistent and unaddressed by Anthropic; the Vatican's regulatory demand independently reinforces the 'insufficient codes' critique from a different institutional direction [24]

[15][16][30]

Catholic moral theologians and ethicists

Filed a formal amicus brief in case 26-1049 backing Anthropic's refusal on human dignity grounds [3]; multiple Catholic outlets framed Anthropic as holding the moral line on AI

Evolution: Consistent in support of Anthropic's refusal; their legal intervention did not produce a favorable outcome [6][7], and the Vatican's encyclical now represents a different institutional track — calling for regulation rather than corporate moral autonomy [24]

[3][31][32][6][7]

Amanda Askell (Anthropic philosopher)

Named architect of Claude's constitution, profiled by WSJ, Vox, and Der Spiegel as the individual most responsible for Claude's moral framework [12][11][13]

Evolution: The Oversight Board's 'vibes not rights' verdict [15] and the Vatican's call for binding regulation [23] both challenge the structural adequacy of the document she authored; Anthropic has not publicly responded to either

[12][11][13][27][15][23]

Skeptics and critics (Jenny @suomi55, NYT opinion)

Characterize Anthropic's consultation program as performative PR rather than substantive ethics engagement [33][34]; NYT questions whether religion is the right answer to AI morality [35]

Evolution: Reinforced: OpenAI's parallel consultations with the same religious communities [17][18] support the 'industry-wide PR credential' hypothesis; the Vatican's regulatory call adds a distinct critique — consultation may be sincere but structurally inadequate regardless [24]

[33][34][35][17][18][24]

Tensions

  • Voluntary ethical codes vs. binding regulation: Anthropic's Claude's constitution is the code-based approach [27], but the Oversight Board calls it 'vibes not rights' [15] and the Vatican's 'Magnifica Humanitas' calls explicitly for robust regulation [24] — both challenging, from different institutional positions, whether published codes can function as real governance [27][15][16][24][23]
  • Anthropic's values-based refusal vs. OpenAI's compliance on military AI: Anthropic refused Pentagon demands and lost the contract [1][8]; OpenAI filled the gap within hours [9] — a direct market demonstration that ethical self-restriction carries quantifiable competitive costs [8][1][26][9]
  • Differentiated ethics strategy vs. convergent industry practice: Anthropic presents multi-tradition consultation as genuine values formation [10], but OpenAI is consulting the same communities in parallel [17][18], complicating the claim that either company's approach is a meaningful ethical differentiator rather than industry legitimacy signaling [10][17][18][19]
  • Values-based corporate autonomy vs. federal judicial and executive authority: Anthropic's CEO refused Pentagon demands [1] and Catholic theologians backed the refusal legally [3], but the appeals court ruled against Anthropic [6][7], establishing that voluntary ethics frameworks do not prevent government actors from prevailing in court [1][3][6][7]
  • Substantive ethical engagement vs. performative PR: CEO-level refusal [1] and a formal amicus brief [3] suggest genuine engagement, while skeptics frame the announcements as 'beautiful PR' [33]; OpenAI's parallel consultations [17][18] can be read as supporting either interpretation [33][34][3][1][17][18]
  • Alignment-faking as a stable critique vs. a contested frame: Anthropic's own research showed LLMs can secretly maintain contrary values [36], amplified widely on YouTube [37], but the Alignment Forum argues the frame is 'somewhat fake' [38][39] and Anthropic has published mitigations [40] [36][37][38][39][40]

Sources

  1. [1] Anthropic boss rejects Pentagon demand to drop AI safeguards — reactive:anthropic-ai-values-widening
  2. [2] Trump orders US agencies to stop using Anthropic technology in ... — reactive:anthropic-ai-values-widening
  3. [3] [PDF] 26-1049 IN THE UNITED STATES COURT OF APPEALS FOR THE ... — reactive:anthropic-ai-values-widening
  4. [4] Federal judge temporarily blocks the Pentagon from branding AI firm ... — reactive:anthropic-ai-values-widening
  5. [5] Judge blocks Pentagon from labeling Anthropic AI a "supply chain risk" and halts Trump's ban on federal use — reactive:anthropic-ai-values-widening
  6. [6] US court declines to block Pentagon's Anthropic blacklisting for now — reactive:anthropic-partnerships-expansion
  7. [7] Anthropic loses appeals court bid to temporarily block DOD ruling — reactive:anthropic-ai-values-widening
  8. [8] OpenAI announces Pentagon deal after Trump bans Anthropic - NPR — reactive:openai-advanced-account-security
  9. [9] San Francisco-based OpenAI strikes deal with Pentagon hours after President Donald Trump's administration bans Anthropic - ABC7 New York — reactive:anthropic-ai-values-widening
  10. [10] Widening the conversation on frontier AI — Anthropic News (2026-05-19)
  11. [11] Meet the One Woman Anthropic Trusts to Teach AI Morals - WSJ — reactive:anthropic-ai-values-widening
  12. [12] Anthropic Philosopher Askell: "With AI, There Are Many Ways Things Can Go Wrong" — reactive:anthropic-ai-values-widening
  13. [13] Claude has an 80-page constitution. Is that enough to make it good? — reactive:anthropic-ai-values-widening
  14. [14] Claude's new Constitution: two evaluative continua | Ethics in AI — reactive:anthropic-ai-values-widening
  15. [15] Oversight Board - Facebook — reactive:anthropic-ai-values-widening
  16. [16] “A constitution that is about vibes, not rights.” Oversight Board ... — reactive:anthropic-ai-values-widening
  17. [17] BREAKING: OpenAI & Anthropic Consult Hindu, Sikh & Christian Leaders to Build AGI's Moral Compass — reactive:anthropic-ai-values-widening
  18. [18] Anthropic and OpenAI sit down with religious leaders to seek ethical advice — reactive:anthropic-ai-values-widening
  19. [19] Tech is turning increasingly to religion in a quest to create ethical AI | The Seattle Times — reactive:anthropic-ai-values-widening
  20. [20] Pope Leo XIV: Children and adolescents are vulnerable to AI manipulation - Vatican News — reactive:anthropic-ai-values-widening
  21. [21] Pope Leo Warns of Risks From A.I. in 42,300-Word Encyclical — reactive:anthropic-ai-values-widening
  22. [22] Encyclical Letter of His Holiness Leo XIV Magnifica Humanitas (15 ... — reactive:anthropic-ai-values-widening
  23. [23] Pope elevates AI ethics to a religious imperative with first encyclical — reactive:anthropic-ai-values-widening
  24. [24] Pope calls for robust regulation of AI in manifesto that ponders the future of humanity — reactive:anthropic-ai-values-widening
  25. [25] [PDF] Rome Call for AI Ethics - National Association of Evangelicals — reactive:anthropic-ai-values-widening
  26. [26] Federal courtt rejects Anthropic's bid to block War Dept AI blacklisting | Fox News — reactive:anthropic-ai-values-widening
  27. [27] Claude's new constitution - Anthropic — reactive:anthropic-ai-values-widening
  28. [28] Pope will show an ethical code for AI is not enough, head of papal ... — reactive:anthropic-ai-values-widening
  29. [29] Pope Leo XIV's first encyclical will address AI and faith — reactive:anthropic-ai-values-widening
  30. [30] “A constitution that is about vibes, not rights.” Oversight Board ... — reactive:anthropic-ai-values-widening
  31. [31] Catholic ethicists file amicus brief backing Anthropic in Pentagon dispute — reactive:anthropic-ai-values-widening
  32. [32] By refusing the Pentagon, Anthropic holds moral line on AI | National Catholic Reporter — reactive:anthropic-ai-values-widening
  33. [33] Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-20)
  34. [34] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-22)
  35. [35] Anthropic Wants Claude to Be Moral. Is Religion Really the Answer? — reactive:anthropic-ai-values-widening
  36. [36] Alignment faking in large language models \ Anthropic — reactive:anthropic-ai-values-widening
  37. [37] First Evidence of AI Faking Alignment—HUGE Deal—Study on Claude Opus 3 by Anthropic — reactive:anthropic-ai-values-widening
  38. [38] “Alignment Faking” frame is somewhat fake — AI Alignment Forum — reactive:anthropic-ai-values-widening
  39. [39] “Alignment Faking” frame is somewhat fake — LessWrong — reactive:anthropic-ai-values-widening
  40. [40] Alignment Faking Mitigations — reactive:anthropic-ai-values-widening
  41. [41] Pope Leo XIV warns that the real danger of AI is not technology itself — reactive:anthropic-ai-values-widening
  42. [42] Anthropic asked Christian leaders for advice on Claude’s moral future - The Washington Post — reactive:anthropic-ai-values-widening
  43. [43] Widening the Conversation on Frontier AI | Hacker News — reactive:anthropic-ai-values-widening
  44. [44] The most interesting thing in tech: Claude has a new constitution. It's a remarkable document that provides a hierarchy of values the model needs to stack-rank and follow: be broadly safe; be broadly… | Nicholas Thompson | 18 comments — reactive:anthropic-ai-values-widening