Version 3

Pope Leo XIV's 'Magnifica Humanitas,' signed May 15 and released May 25, 2026, is the Catholic Church's first encyclical devoted to artificial intelligence, framing AI development as a choice between Babel (power concentration, cultural erasure) and Jerusalem (shared governance, community ownership).[2] Anthropic co-founder Christopher Olah spoke at the Vatican launch, was publicly thanked by the Pope, and Anthropic subsequently published his full remarks — in which Olah acknowledges that frontier AI labs 'operate inside a set of incentives and constraints that can sometimes conflict with doing the right thing' and discloses that Anthropic's interpretability research found AI internal states that 'functionally mirror joy, satisfaction, fear, grief, and unease.'[9][8] AFP's 'disarm AI' wire-service shorthand now dominates global coverage, circulating far more widely than the Babel/Jerusalem metaphor.[18] Indian diplomatic missions are publicly aligning the encyclical with PM Modi's AI governance vision, signaling geopolitical traction beyond Catholic institutional networks.[23][21]

Olah's published remarks — candid about Anthropic's structural incentive conflicts and explicitly inviting external moral scrutiny of his own lab — reframe the corporate-capture controversy into an interpretive question: is this genuine epistemic humility or a more sophisticated form of legitimacy-seeking? If the encyclical succeeds in anchoring AI governance vocabulary across Catholic networks, global diplomatic circuits, and European regulatory platforms simultaneously, it will have achieved something no tech-company policy paper could — moral authority insulated from commercial interest.

Pope Leo XIV signed 'Magnifica Humanitas' on May 15, 2026, and released it publicly on May 25 — the first papal encyclical devoted entirely to artificial intelligence, running to approximately 42,300 words.[1][2][3] The document's organizing metaphor is binary: AI can become Babel, producing uniformity, erasing local culture, and concentrating decision-making among elites, or Jerusalem, a shared rebuilding where every community contributes to and owns a piece of the future.[4] Its foundational premise is that technology 'is never neutral' because it absorbs the values of those who devise, finance, regulate, and deploy it. On practical governance, the document demands explainable reasoning in any AI system affecting jobs, credit, or access to services, imposes social obligations on data governance including worker protections, and calls for regulation preventing monopolization.[4][5] Commentators have widely noted the historical echo: Pope Leo XIII published Rerum Novarum in 1891 to address industrialization's human costs; Leo XIV's encyclical performs the same function for the AI age.[6][7]

The defining controversy of the encyclical's reception concerns Anthropic's role. Christopher Olah, co-founder of Anthropic, spoke at the Vatican launch event and was publicly thanked by Pope Leo XIV.[8] Anthropic subsequently published his full remarks — unusual in their candor for a tech executive. Olah acknowledged that 'every frontier AI lab — including Anthropic — operates inside a set of incentives and constraints that can sometimes conflict with doing the right thing,' described AI models as grown rather than engineered and therefore 'mysterious even to those of us who train them,' and disclosed that Anthropic's internal interpretability research found structures 'that functionally mirror joy, satisfaction, fear, grief, and unease — I don't know what that means, but I think it warrants ongoing discernment.'[9] Olah explicitly framed religious communities as 'necessary external critics whose values cannot be bent by market and competitive pressures,' positioning the Church as a check on his own lab. WIRED investigated why the Vatican invited Anthropic to the presentation;[10] the National Catholic Reporter framed the encounter as a call for 'church-tech ethics partnership.'[11] Tech commentator Corey Quinn, amplified by Simon Willison, had charged that Anthropic's influence constituted 'the single greatest act of vendor lobbying I have ever seen.'[12] Olah's published candor has complicated that charge without fully resolving it.

The encyclical's institutional reception has been broad and geographically wide. Within the Catholic ecosystem, the USCCB, Caritas International, Ascension Press, Aquinas College, Boston Catholic, and Opus Dei have each produced guides, hubs, and academic reflections, spanning the ideological range of the Church.[13][14][15][16][17] AFP's global wire service amplified 'Pope Leo XIV calls for AI to be disarmed' and 'AI Can Never Be Human,' phrases now circulating in dozens of languages and dominating mainstream coverage far more than the Babel/Jerusalem framing.[18][19][20] Indian diplomatic missions — including postings in Georgia, Sydney, and Rajshahi — have begun publicly drawing parallels between the encyclical and PM Modi's AI vision, framing both as 'ethics before efficiency, human-centric AI.'[21][22][23] The EU's Futurium platform published a secular response titled 'Two Magnificences,' suggesting European regulatory circles are engaging with the document.[24]

Two credibility challenges remain unresolved. The Verge identified textual anomalies — including a suspicious pangram — suggesting AI tools may have drafted portions of the document warning against AI, a paradox the Vatican has not publicly addressed.[25] Social media critics have charged that the encyclical 'looks more like a marketing campaign than an encyclical'[26] and that 'many parts were evidently written by AI.'[27] A Protestant appraisal and critical coverage from Transformer News represent cross-denominational and secular skepticism.[28][29] The encyclical's stakes are amplified by Goldman Sachs's projection of $800 billion in annual AI infrastructure spending by end of 2026 and the DHS/FBI's compilation of files on 'anti-technology extremists' — a backdrop in which the Pope's binary test carries particular weight.[4]