The Information Machine

Anthropic's Push to Broaden AI Values Input · history

Version 6

2026-05-25 06:51 UTC · 134 items

What

Anthropic's dispute with the U.S. government now has a clearer legal sequence and a concrete market consequence. A federal judge temporarily blocked the Pentagon from blacklisting Anthropic on March 26 [3][4], but a federal appeals court reversed that on April 8 [5][6]. NPR reported that the ban was a Trump administration action, and OpenAI announced a Pentagon AI deal to fill the contract gap Anthropic's refusal created [2]. On the values side, both Anthropic and OpenAI are now known to be consulting Hindu, Sikh, and Christian religious leaders [20][19], while the Oversight Board's dismissal of Claude's constitution as 'about vibes, not rights' [13] remains publicly unaddressed and a Medium piece argues AI alignment is irrelevant to safety altogether [26].

Why it matters

OpenAI's capture of the Pentagon contract Anthropic refused makes the market cost of ethical self-regulation concrete: values-based refusals are only sustainable when a company can absorb losses a competitor will not. The convergence of both major AI companies on the same religious consultation practice [20][19] simultaneously complicates the claim that either firm's approach represents differentiated ethics work, raising the question of whether religious consultation has become industry-standard legitimacy signaling rather than genuine values formation.

Open questions

  • With the appeals court ruling against it [5][6] and OpenAI claiming the Pentagon contract [2], what are Anthropic's remaining legal and commercial options — and does the revenue loss make future values-based refusals of government demands harder to sustain?

  • Both Anthropic and OpenAI are consulting the same religious communities [20][19] — does this convergence make religious consultation a form of industry credential rather than a differentiating ethics strategy, and what would meaningful differentiation look like?

  • Will Anthropic revise Claude's constitution in response to the Oversight Board's 'vibes not rights' verdict [13][14], and if rights provisions were added, what institution would hold authority to enforce them?

  • A Medium argument that 'AI alignment is irrelevant to AI safety' [26] strikes at the foundational premise of the consultation initiative — does Anthropic's dual-track strategy have a public response to this class of critique, and does the court's adverse ruling add empirical weight to it?

Narrative

Anthropic's conflict with the U.S. government spans a documented legal arc. The dispute originated when the Pentagon demanded Anthropic drop its AI safeguards as a condition for military use; the company's CEO explicitly refused, and the DoD blacklisted Anthropic in response [1]. The conflict carried executive-branch framing: NPR reported that the Trump administration banned Anthropic from federal use, and OpenAI announced a Pentagon deal to fill the resulting AI contract gap [2]. On March 26, 2026, a federal judge temporarily blocked the Pentagon from branding Anthropic a 'supply chain risk' and halted the federal ban [3][4]. That reprieve was brief: a federal appeals court overturned the temporary block on April 8, declining to extend it [5][6]. Catholic moral theologians had filed a formal amicus brief in case 26-1049 on March 16, backing Anthropic's refusal on 'human dignity' grounds [7], but the adverse ruling stands. A Wikipedia article [8] and a TechPolicy.Press timeline [9] now document the dispute as a reference case.

Claude's published constitution has attracted a widening circle of institutional responses. Nicholas Thompson called it 'the most interesting thing in tech' on LinkedIn, describing its values hierarchy as a novel governance artifact [10]. BISI published a governance analysis framing it as a model-governance precedent [11], and the LessWrong community opened a dedicated discussion thread [12]. The most pointed institutional critique came from the Oversight Board, whose member Suzanne N. issued the verdict that it is 'a constitution that is about vibes, not rights' [13][14] — a direct challenge to whether the document carries enforceable weight rather than aspirational framing. Oxford AI Ethics offered a formal academic analysis through 'two evaluative continua' [15], and Amanda Askell, the philosopher named as the document's primary architect, was profiled by the WSJ, Vox, and Der Spiegel [16][17][18].

Anthropic’s practice of consulting religious and humanistic communities for AI values input is no longer singular. The Decoder and a YouTube explainer both reported that OpenAI is also sitting down with Hindu, Sikh, and Christian leaders for ethical guidance [19][20], placing both major AI labs in the same practice simultaneously. A postsecular-perspective analysis on LinkedIn offered academic framing for the convergence [21], and Let's Data Science reported on Anthropic's faith-leader meetings alongside similar industry initiatives [22]. At the highest level of institutional Catholicism, Pope Leo XIV addressed AI in his first papal remarks — referenced in current media as relevant context for the religious AI debate [23]. The Seattle Times had already situated Anthropic's outreach within a broader tech-to-religion industry turn [24]; OpenAI's parallel consultations strengthen that framing.

The theoretical debate about what consultation work actually accomplishes runs alongside the legal and institutional story. Paul Christiano has argued that technical and social AI safety approaches are complementary rather than substitutes [25], but a Medium piece published in May 2026 took a starker position: that AI alignment is 'irrelevant to AI safety' altogether [26], a claim that would undercut the premise of values consultation if accepted. The Alignment Forum contested the 'alignment faking' critique of consultation, arguing the frame is 'somewhat fake' [27][28], and Anthropic has published mitigations to alignment-faking behavior [29]. Whether humanistic values consultation is a necessary complement to technical safety work, a legitimate substitute, or irrelevant to safety remains live and unresolved.

Timeline

  • 2026-02-27: Trump administration bans Anthropic from Pentagon use; OpenAI announces a Pentagon AI deal to fill the resulting contract gap [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 [7]
  • 2026-03-19: Washington Post reports Catholic thinkers object to Pentagon AI demands on 'human dignity' grounds [60]
  • 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 [3][4]
  • 2026-04-08: Federal appeals court overturns the temporary block, declining to prevent the Pentagon's blacklisting of Anthropic [5][6]
  • 2026-04-11: Washington Post reports that Anthropic consulted Christian leaders for advice on Claude's moral future [59]
  • 2026-04-20: New York Times publishes opinion piece 'Anthropic Wants Claude to Be Moral. Is Religion Really the Answer?' questioning the initiative's approach [40]
  • 2026-05-19: Anthropic publishes 'Widening the conversation on frontier AI,' describing dialogues with 15+ religious and cross-cultural traditions and disclosing the ethical-reminder tool experiment [63]
  • 2026-05-20: Rohan Paul amplifies the Anthropic post on X; Jenny (@suomi55) posts skeptical 'beautiful PR' characterization; multiple accounts amplify both [62][41][66]
  • 2026-05-21: Skeptical 'PR post' framing spreads across more than a dozen accounts; Hacker News thread on the initiative opens [67][42][44][45][46][47][48][49][50][51][52][53][54][55][56]
  • 2026-05-23: WSJ, Vox, and Der Spiegel profiles of Amanda Askell surface; Vox reports Claude's moral framework runs to ~80 pages; Anthropic's alignment-faking mitigations circulate alongside Alignment Forum critique of the 'alignment faking' frame [16][17][18][27][28][29]
  • 2026-05-24: Anthropic publishes Claude's constitution publicly; Oxford AI Ethics and Oversight Board respond with external analyses; Catholic moral theologians' amicus brief circulates widely across Catholic media [31][15][38][32][33][34][35][36]
  • 2026-05-24: BBC reports Anthropic CEO explicitly rejected Pentagon demands; Oversight Board member Suzanne N. characterizes Claude's constitution as 'about vibes, not rights'; Seattle Times frames the initiative as part of a broader industry turn to religion for AI ethics; Nicholas Thompson endorses the constitution on LinkedIn; multiple sources confirm OpenAI is also consulting Hindu, Sikh, and Christian leaders [1][13][24][10][20][19]

Perspectives

Anthropic

Publicly maintained a values-based refusal of Pentagon demands to drop AI safeguards, with the CEO explicitly declining compliance [1]; lost a federal appeals court challenge to block the resulting blacklisting [5][6], with OpenAI filling the Pentagon contract gap [2]; the published Claude constitution has been dismissed by the Oversight Board as 'about vibes, not rights' [13]

Evolution: Compounded setback: the Trump administration ban [2] and appeals court loss [5][6] are now joined by a concrete competitive consequence — OpenAI claiming the Pentagon contract — and the Oversight Board criticism remains unaddressed

[1][30][13][31][2][5][6]

Trump administration / U.S. Department of Defense

The Trump administration banned Anthropic from federal use after the company refused to comply with Pentagon AI demands [2]; the DoD blacklisted Anthropic and prevailed in the federal appeals court when Anthropic challenged the blacklisting [5][6]

Evolution: The political dimension is now named: prior synthesis identified only the DoD as actor; the NPR report [2] frames the ban as a Trump administration decision, adding executive-branch political context to what was previously characterized as a procurement dispute

[2][5][6][1]

OpenAI

Announced a Pentagon deal to fill the AI contract gap left by Anthropic's ban [2]; also independently consulting Hindu, Sikh, and Christian religious leaders for ethical guidance [20][19] — positioning OpenAI as both the direct market beneficiary of Anthropic's values-based refusal and a parallel actor in the religious consultation space

Evolution: Newly named as a significant actor in this thread; OpenAI's dual role — Pentagon contract winner and religious-consultation participant — makes it the structural mirror image of Anthropic on both fronts

[2][20][19]

Catholic moral theologians and ethicists

Filed a formal amicus brief in case 26-1049 backing Anthropic's refusal to comply with Pentagon demands [7]; multiple Catholic outlets framed Anthropic as 'holding the moral line on AI'; Catholic University characterized the dispute as 'Autonomous Weapons vs. Moral Agents'

Evolution: Consistent; the court ruled against Anthropic despite the brief, meaning religious legal intervention did not produce a favorable outcome [5][6]

[7][32][33][34][35][36]

Oversight Board

Member Suzanne N. characterized Claude's published constitution as 'a constitution that is about vibes, not rights' [13][14], issuing a blunt dismissal that goes beyond calling for additional rights mechanisms to a direct indictment of the document's substantive adequacy

Evolution: Facebook post [13] confirms the quote and names member Suzanne N. as source; stance unchanged and remains publicly unaddressed by Anthropic

[13][14][37][38]

Amanda Askell (Anthropic philosopher)

Named architect of the now-public Claude constitution; profiled by WSJ, Vox, and Der Spiegel as the individual most responsible for Claude's moral framework; her moral precepts noted as having 'gone viral' in January 2026 [39]

Evolution: Broader institutional reception of the constitution — Nicholas Thompson's endorsement [10], BISI analysis [11], LessWrong thread [12] — adds reach to the document she authored, but the Oversight Board verdict and adverse court ruling affect its institutional standing

[16][17][18][31][39][10]

Nicholas Thompson (LinkedIn)

Called Claude's constitution 'the most interesting thing in tech,' describing its values hierarchy as a novel and significant governance artifact [10]

Evolution: New voice this pass; endorsing perspective provides a prominent non-specialist counterweight to the Oversight Board's dismissal

[10]

Oxford AI Ethics (University of Oxford)

Published a formal academic analysis of Claude's constitution, framing it through 'two evaluative continua' as an analytical lens [15]

Evolution: Consistent

[15]

BISI

Published a governance analysis of Claude's constitution framing it as a model-governance precedent with implications for AI alignment and the future of model governance [11]

Evolution: New voice this pass; adds a distinct institutional governance framing to the constitution debate

[11]

Pope Leo XIV / Vatican

Addressed AI in his first papal remarks — cited in current media as relevant context for the religious AI debate [23]

Evolution: New voice surfaced this pass; the papal address dates to May 2025 but its relevance to the current debate is being amplified now as the religious-AI engagement story widens

[23]

Seattle Times

Reports that the tech industry broadly is turning to religion in a quest for ethical AI guidance [24], situating Anthropic's initiative within an industry-wide pattern rather than treating it as isolated

Evolution: Reinforced this pass: OpenAI's parallel religious consultations [20][19] provide empirical support for the industry-pattern framing

[24][20][19]

Paul Christiano (AI safety researcher)

Argues that technical and social approaches to AI safety are distinct but complementary, not substitutes [25]

Evolution: Consistent; the Medium piece arguing alignment is irrelevant to safety [26] now represents a direct challenge to this complementarity thesis

[25]

Medium / 'AI Alignment Is Irrelevant to AI Safety'

Argues that AI alignment work is irrelevant to the AI safety problem altogether [26], a position that would undercut the foundational premise of both technical alignment research and values consultation if accepted

Evolution: New voice this pass; represents a starker critique than either the NYT's question about religion or the Alignment Forum's dispute over the alignment-faking frame

[26]

New York Times (opinion)

Questions whether religious consultation is the right answer to the problem of AI morality [40]

Evolution: Consistent

[40]

Jenny (@suomi55) and amplifiers

Skeptical and dismissive — characterizes Anthropic's announcement as performative PR rather than substantive engagement [41][42]

Evolution: Consistent; OpenAI's parallel consultations [20][19] could be read as supporting the PR-framing hypothesis if both companies are doing it simultaneously

[41][42][43][44][45][46][47][48][49][50][51][52][53][54][55][56][57]

Alignment Forum / LessWrong community

Contests the 'alignment faking' critique of consultation, arguing the frame is 'somewhat fake' [27][28]; opened a dedicated discussion thread on Claude's new constitution [12]

Evolution: Expanded engagement: the LessWrong thread [12] adds direct community analysis of the constitution document to the prior alignment-faking debate

[27][28][58][12]

Washington Post

Reported Christian leader consultations in April 2026 before Anthropic's public announcement; also reported Catholic objections to Pentagon AI in March 2026 — consistently framing religious engagement with AI as factual news

Evolution: Consistent

[59][60][61]

Rohan Paul (@rohanpaul_ai)

Amplifies and endorses Anthropic's framing that frontier AI development requires scholars, philosophers, clergy, and civic thinkers as essential contributors

Evolution: Consistent

[62]

Tensions

  • Anthropic's refusal vs. OpenAI's compliance on military AI: Anthropic refused Pentagon demands on ethical grounds and lost a major government contract [2][1]; OpenAI immediately filled that gap [2] — creating a direct market demonstration that values-based refusals have quantifiable competitive costs when a less restrictive competitor is available. [2][1][30]
  • Values-based corporate refusal vs. federal judicial and executive authority: Anthropic's CEO explicitly refused Pentagon demands [1], a federal judge temporarily backed Anthropic [3][4], and Catholic theologians filed a formal amicus brief in support [7] — but a federal appeals court overturned the temporary block [5][6] and the Trump administration's ban stands, establishing that voluntary ethics frameworks do not prevent government actors from prevailing in court. [1][3][4][7][5][6][2]
  • Differentiated ethics strategy vs. convergent industry practice: Anthropic presents multi-tradition consultation as genuine values formation [63], but OpenAI is conducting parallel religious consultations with the same communities [20][19] — complicating the claim that either company's approach is a meaningful ethical differentiator rather than industry-standard legitimacy signaling. [63][20][19][24]
  • Substantive engagement vs. performative PR: Catholic ethicists' amicus brief [7] and CEO-level refusal [1] provide evidence of substantive engagement, while skeptics characterize Anthropic's announcements as 'beautiful PR' [41][42]; the simultaneous discovery that OpenAI runs parallel consultations [20][19] can be read as supporting either side — genuine industry norm or industry-wide PR strategy. [63][41][42][7][1][20][19]
  • Self-governance vs. external oversight: Anthropic published Claude's constitution as its authoritative values document [31], while the Oversight Board dismissed it as 'about vibes, not rights' [13][14] — a direct challenge to whether Anthropic can adequately govern its own values framework without external enforcement. [31][13][14][37][38]
  • Religious and humanistic consultation vs. technical alignment relevance: The NYT questions whether religion is the right answer to AI morality [40], while a Medium piece goes further and argues alignment is irrelevant to safety altogether [26]; Paul Christiano frames technical and social approaches as complementary [25], but the court's adverse ruling against Anthropic's values-based legal position challenges the practical enforceability of the humanistic track against state actors. [40][26][25][30]
  • Singular moral authority vs. plural governance: The WSJ frames Amanda Askell as 'the one woman Anthropic trusts to teach AI morals' [17], concentrating moral authority in a single named individual — in tension with Anthropic's stated goal of broad multi-tradition consultation where no single framework dominates [63]. [17][63]
  • Alignment-faking as a stable critique vs. a contested frame: Anthropic's own research showed LLMs can secretly maintain contrary values [64][65], cited as undermining the consultation premise — but the Alignment Forum argues the 'alignment faking' frame is itself 'somewhat fake' [27][28], and Anthropic has published mitigations [29], leaving the technical backstory disputed. [64][65][27][28][29][63]

Sources

  1. [1] Anthropic boss rejects Pentagon demand to drop AI safeguards — reactive:anthropic-ai-values-widening
  2. [2] OpenAI announces Pentagon deal after Trump bans Anthropic - NPR — reactive:openai-advanced-account-security
  3. [3] Federal judge temporarily blocks the Pentagon from branding AI firm ... — reactive:anthropic-ai-values-widening
  4. [4] Judge blocks Pentagon from labeling Anthropic AI a "supply chain risk" and halts Trump's ban on federal use — reactive:anthropic-ai-values-widening
  5. [5] US court declines to block Pentagon's Anthropic blacklisting for now — reactive:anthropic-partnerships-expansion
  6. [6] Anthropic loses appeals court bid to temporarily block DOD ruling — reactive:anthropic-ai-values-widening
  7. [7] [PDF] 26-1049 IN THE UNITED STATES COURT OF APPEALS FOR THE ... — reactive:anthropic-ai-values-widening
  8. [8] Anthropic–United States Department of Defense dispute - Wikipedia — reactive:anthropic-partnerships-expansion
  9. [9] A Timeline of the Anthropic-Pentagon Dispute | TechPolicy.Press — reactive:openai-financial-strategy
  10. [10] 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
  11. [11] Claude's New Constitution: AI Alignment, Ethics, and the Future of ... — reactive:anthropic-ai-values-widening
  12. [12] Claude's new constitution — LessWrong — reactive:anthropic-ai-values-widening
  13. [13] Oversight Board - Facebook — reactive:anthropic-ai-values-widening
  14. [14] “A constitution that is about vibes, not rights.” Oversight Board ... — reactive:anthropic-ai-values-widening
  15. [15] Claude's new Constitution: two evaluative continua | Ethics in AI — reactive:anthropic-ai-values-widening
  16. [16] Anthropic Philosopher Askell: "With AI, There Are Many Ways Things Can Go Wrong" — reactive:anthropic-ai-values-widening
  17. [17] Meet the One Woman Anthropic Trusts to Teach AI Morals - WSJ — reactive:anthropic-ai-values-widening
  18. [18] Claude has an 80-page constitution. Is that enough to make it good? — reactive:anthropic-ai-values-widening
  19. [19] Anthropic and OpenAI sit down with religious leaders to seek ethical advice — reactive:anthropic-ai-values-widening
  20. [20] BREAKING: OpenAI & Anthropic Consult Hindu, Sikh & Christian Leaders to Build AGI's Moral Compass — reactive:anthropic-ai-values-widening
  21. [21] Faith and the Ethics of Artificial Intelligence: A Postsecular Perspective — reactive:anthropic-ai-values-widening
  22. [22] Anthropic and Faith Leaders Meet on AI Ethics | Let's Data Science — reactive:anthropic-ai-values-widening
  23. [23] Last May, Pope Leo XIV called out AI in his first address ... - Instagram — reactive:anthropic-ai-values-widening
  24. [24] Tech is turning increasingly to religion in a quest to create ethical AI | The Seattle Times — reactive:anthropic-ai-values-widening
  25. [25] Technical and social approaches to AI safety | by Paul Christiano — reactive:anthropic-ai-values-widening
  26. [26] AI Alignment Is Irrelevant to AI Safety | Write A Catalyst | May, 2026 — reactive:anthropic-ai-values-widening
  27. [27] “Alignment Faking” frame is somewhat fake — AI Alignment Forum — reactive:anthropic-ai-values-widening
  28. [28] “Alignment Faking” frame is somewhat fake — LessWrong — reactive:anthropic-ai-values-widening
  29. [29] Alignment Faking Mitigations — reactive:anthropic-ai-values-widening
  30. [30] Federal courtt rejects Anthropic's bid to block War Dept AI blacklisting | Fox News — reactive:anthropic-ai-values-widening
  31. [31] Claude's new constitution - Anthropic — reactive:anthropic-ai-values-widening
  32. [32] Catholic ethicists file amicus brief backing Anthropic in Pentagon dispute — reactive:anthropic-ai-values-widening
  33. [33] By refusing the Pentagon, Anthropic holds moral line on AI | National Catholic Reporter — reactive:anthropic-ai-values-widening
  34. [34] Catholic moral theologians, ethicists back Anthropic in ... — reactive:anthropic-ai-values-widening
  35. [35] Anthropic fight with US Pentagon amid Iran war… — reactive:anthropic-ai-values-widening
  36. [36] Autonomous Weapons vs. Moral Agents: A Theologian Discusses the Anthropic Case | Catholic University — reactive:anthropic-ai-values-widening
  37. [37] “A constitution that is about vibes, not rights.” Oversight Board ... — reactive:anthropic-ai-values-widening
  38. [38] Claude's Constitution Needs a Bill of Rights and Oversight — reactive:anthropic-ai-values-widening
  39. [39] This January, a set of moral precepts for Anthropic's chatbot, Claude ... — reactive:anthropic-ai-values-widening
  40. [40] Anthropic Wants Claude to Be Moral. Is Religion Really the Answer? — reactive:anthropic-ai-values-widening
  41. [41] Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-20)
  42. [42] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-22)
  43. [43] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-22)
  44. [44] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  45. [45] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  46. [46] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  47. [47] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  48. [48] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  49. [49] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  50. [50] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  51. [51] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  52. [52] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  53. [53] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  54. [54] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  55. [55] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  56. [56] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-21)
  57. [57] RT @suomi55: Anthropic just dropped another beautiful PR post: — reactive:anthropic-ai-values-widening (2026-05-20)
  58. [58] ALIGNMENT FAKING IN LARGE LANGUAGE MODELS - Medium — reactive:claude-evaluation-awareness
  59. [59] Anthropic asked Christian leaders for advice on Claude’s moral future - The Washington Post — reactive:anthropic-ai-values-widening
  60. [60] To Catholic thinkers, Pentagon’s AI demands violate ‘human dignity’ - The Washington Post — reactive:anthropic-ai-values-widening
  61. [61] Anthropic, an AI company, hosted Christian religious leaders at its ... — reactive:anthropic-ai-values-widening
  62. [62] Anthropic's new study says frontier AI needs input from scholars, philosophers, clergy, and civic thinkers because model… — Rohan Paul Twitter (2026-05-20)
  63. [63] Widening the conversation on frontier AI — Anthropic News (2026-05-19)
  64. [64] Alignment faking in large language models \ Anthropic — reactive:anthropic-ai-values-widening
  65. [65] New Anthropic study: LLMs can secretly transmit personality traits ... — reactive:anthropic-ai-values-widening
  66. [66] Anthropic is expanding the conversation around frontier AI by ... — reactive:anthropic-ai-values-widening
  67. [67] Widening the Conversation on Frontier AI | Hacker News — reactive:anthropic-ai-values-widening