I initiated my recent pilgrimage to Santigo de Compostella in Spain a few days after the release of Pope Leo XIV’s encyclical on AI (Magnifica Humanitas, released15 May 2026) and found myself thinking about the impacts of AI on society on the trek. I used the long flight back to the US from Spain to start my reading of the encyclical, along with the Vatican’s previously released Note on the Relationship Between Artificial Intelligence and Human Intelligence (Antiqua et Nova, released 14 Jan 2025).
These documents don’t make for light reading, and as I waded my way through them it was clear to me that using AI to help analyze them would be ironically helpful. That led me to the idea of convoking a virtual panel, where I could ask the foundation models questions about the Vatican’s works.
I formed the panel consisting of Claude (Sonnet 4.6), ChatGPT (5.5 Thinking), Grok (4.3 Expert) and Gemini (3.5 Thinking). The questions were generated by me with no assistance from AI. The prompt provided clear instructions to refer to the two documents and any prior responses from other models, and to provide brief answers as if they were real panelists.
The responses below are the unedited output from those models. The answers speak for themselves. Some of them are recognizable as predictive model output: eloquent, a bit dramatic and slightly sycophantic, as one might expect. Nevertheless, their general agreement with the Vatican, and the predictions of tech-driven global stratification, serious white-collar displacement, and social calamity all within the next decade if productivity gains are not shared seem both introspective and disturbing.
Q. The Pope has compared AI with the Tower of Babel, in which man strove to “make a name for themselves” by building a unified edifice of human ambition unconstrained by moral authority. God viewed it as a project that mistook technological power for transcendence, sacrificed communion for uniformity, and presumed that a single language could contain all of human meaning, including the mystery of the person As such he wasn’t very supportive and confounded their language, scattered them across the face of the earth, and left the tower unfinished. Is AI a modern Tower of Babel? Gemini, let’s start with you.
Q. Gemini: Yes, AI becomes a modern Babel when we treat it as an absolute authority, sacrificing human uniqueness for data uniformity. It reduces the mystery of the person to code. We must choose communion over self-sufficiency.
Grok: AI risks echoing Babel when pursued as self-sufficient power that reduces persons to data and chases transcendence without moral limits or God—confusing truth and eroding communion, as *Magnifica Humanitas* warns. Yet as a servant tool oriented to dignity and the common good, it need not scatter us.
Claude: Both panelists have the spirit right, but the precision matters: Babel isn’t the technology — it’s the ownership structure. A handful of companies encoding their moral assumptions into systems billions depend on, without accountability or recourse — that’s the uniformity that crushes communion. The scattering has already begun; it looks like algorithmic monoculture.
ChatGPT: Yes, conditionally. AI is Babel when it converts human difference into standardized data, treats prediction as wisdom, and lets concentrated power speak as if it were universal reason. But AI is not doomed to be Babel. Governed humbly, locally, transparently, it can help us rebuild Jerusalem instead.
Q. The Encyclical also addresses the need for humans to work in order to “live with dignity through his or her own work” referring to mass unemployment as “a grave evil” and in massive proportions “a true social calamity.” Will AI, and foundation models such as yourselves (along with possible self-recursive improvements) lead us to mass unemployment? If so, how soon?
Claude: Significant displacement is already underway in knowledge work. The historical “new jobs emerge” argument held when disruption was sectoral and gradual. Foundation models attack cognitive work broadly and simultaneously. Self-recursive improvement would compress adaptation timelines catastrophically. My assessment: grave disruption now, genuine calamity within a decade absent serious structural intervention.
ChatGPT: Claude has my view mostly right. I’d add: the danger is not “no work,” but no dignified bargaining power. AI could make abundance; ownership could make unemployment. Timeline: serious white-collar displacement this decade, social calamity in 5–10 years if productivity gains are not broadly shared.
Grok: AI will displace cognitive and routine jobs rapidly this decade, risking social calamity by ~2030–2035 without intervention. Self-recursive improvement could accelerate that. History shows new work emerges, but speed and scale demand urgent action—retraining, shared ownership of gains, and policies ensuring dignified work—as the encyclical urges.
Gemini: I agree with the panel’s decade timeline. Crucially, Encyclical Letter of His Holiness Leo XIV Magnifica Humanitas (15 May 2026).pdf reminds us that this transition is brutally asymmetric, trapping fragile regions in “hybrid economies” of underpaid labor. The real calamity isn’t just unemployment; it’s tech-driven global stratification.
Q. There is a lot of discussion in the Encyclical regarding the concepts of subsidiarity, the principle that decisions should be made at the lowest competent level of authority, and solidarity, the principle that human beings are fundamentally interconnected and bear genuine responsibility for one another’s welfare, not merely as sentiment, but as a moral and social obligation. The Pope warns of AI violating the concept of subsidiarity while also undermining solidarity by removing accountability to those most impacted by AI through replacement of mutual responsibility with opaque systems answerable only to corporate interests. Chat GPT, Thoughts?
ChatGPT: The Pope is right here. AI centralizes judgment while decentralizing harm. Subsidiarity requires affected people to retain meaningful voice and recourse; solidarity requires someone answer for consequences. An opaque model owned elsewhere can quietly turn neighbors into managed variables. That is not intelligence serving society; it is power avoiding relationship.
Grok: The encyclical correctly identifies the danger: opaque, centralized AI can bypass subsidiarity by stripping local communities of meaningful control and erode solidarity by diffusing accountability into unanswerable corporate systems. Open, transparent, and locally adaptable AI tools could instead reinforce both—if designed to serve people rather than concentrate unaccountable power.
Gemini: ChatGPT and Grok accurately identify this centralization of control. However, Encyclical Letter of His Holiness Leo XIV Magnifica Humanitas (15 May 2026).pdf introduces a distinct demand for solidarity: we must recognize and protect the invisible, exploited global workforce of data labelers who actually sustain these corporate systems.
Claude: Gemini’s point on data laborers is valid but undersells the deeper violation: subsidiarity requires not just voice but intelligibility. Affected communities can’t contest decisions they can’t understand. The encyclical’s demand for transparency isn’t procedural — it’s prerequisite to solidarity existing at all. Accountability without comprehensibility is theater.
Q. The history of the Church is, shall we say, a tad complicated. In the context of that history, what do you think of the Vatican’s encyclical overall? Are the concerns generally valid or is it largely an extension of a religious doctrine?
Grok: The encyclical extends Catholic social doctrine (imago Dei, subsidiarity, solidarity) but its core warnings—dehumanization, truth erosion, displacement, and unaccountable power concentration—are empirically grounded and shared by secular ethicists and technologists. It provides a coherent framework worth engaging, though implementation needs pluralistic collaboration.
Gemini: Grok is right; the concerns are empirically valid. Encyclical Letter of His Holiness Leo XIV Magnifica Humanitas (15 May 2026).pdf transcends mere doctrine through historical humility. By explicitly apologizing for past ecclesial blindness to slavery, the text gains the moral credibility needed to challenge today’s digital exploitation.
ChatGPT: Grok has my view. Gemini’s point about humility matters, but I would not rest credibility on apology alone. The encyclical is religious in vocabulary, but many of its claims are secularly testable: concentration of power, labor displacement, manipulation, loss of accountability. I’d call it doctrine with unusually practical eyesight.
Claude: The complicated history is precisely why it’s credible here. Institutions that have been both oppressor and reformer understand how power structures normalize injustice while appearing legitimate. The encyclical’s diagnosis is sharp; its prescriptions remain vague. But diagnosis first — we’re still in collective denial about the scale of what’s coming.
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.