教宗Leo XIV发表了一篇约40,000字的《Magnifica Humanitas》，主张人工智能应“服务人类，而非集中权力”。文章借由教廷通谕的历史角色，强调这类文献往往出现在社会情绪与经济信心的高低点，并被视为极敏锐的时代信号；作者借此判断，AI已成为公共议题的转折点，且争论正从技术问题上升为道德问题。

文中回顾了多次历史对应：Leo XIII在1891年发表《Rerum Novarum》，回应镀金时代与劳工问题；Peter Atwater与Francesca Pereira指出，自1925年以来，9篇高关注度通谕都出现在全球信心的高峰或低谷附近，常与Dow/Gold ratio的极值同步。例子包括1930年的《Casti Connubii》、1968年的《Humanae Vitae》、2009年的《Caritas in Veritate》、2020年的《Fratelli Tutti》与1998年的《Fides et Ratio》，都在高不确定性或强烈乐观/悲观时期发声。

在AI议题上，Leo XIV批评两点：其一是技术由少数公司和缺乏问责的机制主导，其二是AI削弱人的身份、创造力与判断力。文章提到SpaceX、OpenAI、Anthropic今年的IPO计划，象征权力进一步集中；同时社会对大企业与AI的反感上升，包括精英大学毕业典礼上对AI的嘘声，以及围绕Luigi Mangione审判的关注。教宗警告，若不加约束，AI可能放大偏见、排斥弱者，并把“必要的牺牲”合理化，因此需要更严肃、更强烈的道德辩论。

Pope Leo XIV’s roughly 40,000-word *Magnifica Humanitas* argues that AI should serve humanity rather than concentrate power. The piece treats papal encyclicals as unusually sharp indicators of social mood, suggesting that this intervention marks a turning point as AI moves from a technical issue to a broad moral and political one.

The article links major encyclicals to turning points in confidence: nine high-profile encyclicals since 1925 reportedly appeared near extremes in the Dow-to-gold ratio. Historical examples include *Rerum Novarum* (1891), *Casti Connubii* (1930), *Humanae Vitae* (1968), *Caritas in Veritate* (2009), *Fratelli Tutti* (2020), and *Fides et Ratio* (1998), each issued amid major uncertainty, optimism, or social upheaval.

Leo’s core objections are AI’s unaccountable governance and its effects on human identity. He warns that a few dominant firms, amplified by planned IPOs for SpaceX, OpenAI, and Anthropic, could entrench opaque power, while chatbots and algorithmic systems may erode creativity, judgment, and dignity. Against Silicon Valley libertarianism, the encyclical insists that society must intensify the debate over fairness, vulnerability, and whether “necessary sacrifices” are being imposed on the weakest.