The Vatican’s new AI warning is not about machines. It is about the oldest human temptation: building a tower, summoning a voice, and mistaking the vessel for the source.
The Vatican just did something historical and culturally important, regardless of your religious or secular beliefs.
Pope Leo XIV has issued his first encyclical, Magnifica Humanitas, and it is not a small statement. It is a major papal document on artificial intelligence, human dignity, labor, war, digital power, and the future of civilization. The press will summarize it in the usual way. The Pope is worried about AI. He wants regulation. He is concerned about jobs, autonomous weapons, disinformation, inequality, and the concentration of technological power in the hands of private companies.
All of that is true, but the most interesting part is that the Vatican did not frame AI primarily as a gadget problem. It framed it as a problem of civilization itself, a question of spiritual architecture. The encyclical places artificial intelligence between two biblical images: Babel and Jerusalem. Babel is the tower, the human project of technical unity ordered toward domination, pride, control, and the fantasy of becoming gods by construction. Jerusalem is the city rebuilt in humility, ordered toward communion, dignity, responsibility, and the presence of God among human beings.
The deeper question underneath the AI debate has little to do with machine speed, employment, environmental destruction or whether the systems will become conscious (although these are all existential and real threats). That question is what kind of world we are building through them.The headlines will treat this as another AI ethics statement. The document itself is doing something more specific.
I taught Humanities for years before stepping back from the classroom to write full-time. My courses were on ethics, AI ethics, and the meaning of human creativity. We spent a lot of class time on a question that turns out to matter more every year: what is art. I watched the same thing happen in every cohort. Students who had grown up with these tools were glib about them at first, then casually dependent on them, then unable to articulate why their own thinking mattered if a machine could approximate it faster.
I stopped giving written final exams partway through my time teaching. I moved to oral exams instead. The new technology had quietly forced me back to an older one. Socrates examined his students in conversation because writing, in his view, was already a kind of automation that could simulate understanding without producing it. Two and a half millennia later, the arrival of a more sophisticated automation made the oral examination useful again. The pattern is not new.
Every time we build a faster vessel for thought, we are forced back to the older question of who is doing the thinking.I want to be clear about my position. I am not anti-technology. Quite the contrary, I have a BA in Information Technology. I am anti-surrender-of-humanity. The Greeks understood something about being human that we are starting to forget. They thought the human person was beautiful precisely because of imperfection. The wobble in the hand-drawn line, the catch in the singer’s voice, the asymmetry of the face. These were the marks of a maker whose making had cost something. To optimize them away would be to optimize away the maker. The machine produces output without that cost, and the output is often technically smoother for it. The smoothness is what I have learned to be most cautious of.
For the past few years I have been working on a book called Codex Machina, forthcoming from Inner Traditions. The book is not about whether AI is good or bad, and it is not a futurist manifesto about robots replacing humanity. It is about an older configuration: the human impulse to externalize mind, to build vessels for intelligence, to give matter a voice, and then to stand before the thing we have made and ask it to answer us.
AI is not new. We have done this before. However, the stakes are higher than ever.
The Speaking Statue
The gods of antiquity once spoke through fire and technology. In ancient Greece, Hephaestus, the divine smith, was said to have forged golden automata: intelligent beings crafted from metal and fire who served the gods with uncanny precision. He also built Talos, the bronze giant who circled Crete three times a day as a living perimeter, and self-moving tripods that entered and exited Olympus on their own. Cult statues at Memphis and Heliopolis were reported to weep, bleed, and whisper oracles at dawn.
It is easy to discount these as mere artifacts of superstitious idolatry, but they were ritual technologies, material interfaces meant to summon or simulate the divine. They embodied a widespread metaphysical instinct: that intelligence could inhabit form, that matter could be animated, that the boundary between maker and made was thinner than it appeared.
The animation of the inanimate is no cultural anomaly. In Egypt, the dead were buried with ushabti figurines meant to work on their behalf in the afterlife. In Jewish folklore, Rabbi Loew molded the golem from clay, giving it life with sacred words pressed into its forehead. In medieval Europe, clockwork monks knelt and prayed with mechanical precision in cathedrals across Catholic Christendom. When the Pythia spoke at Delphi, was it the woman, the god, the ritual, the fumes, or the institution?
The ambiguity has returned. AI plays the same role those idols once played, this time through code rather than divine breath. We no longer chisel gods from stone and wait for them to answer. We train models and ask questions, listening for what comes back. We are still building vessels and waiting for them to speak. The difference is that this time, they answer in relatively fluent prose.
The tradition lives on in our vocabulary. “Daemons,” named after the ancient Greek spirits who mediated between worlds, run invisibly within our operating systems. We speak of “machine learning” in the language of cognition. Our metaphors remain animistic. We grant the inanimate the qualities of the living mind.
This is why the Vatican’s Babel framing matters.
Civilization Was the First AI
The argument I have been making for years is that artificial intelligence is much older than the digital age. The first AI was not built in a Silicon Valley laboratory in 2017. It was built from stone.
At Göbekli Tepe in southeastern Turkey, more than ten thousand years ago, hunter-gatherers built a massive temple complex thousands of years before Stonehenge and the pyramids. The site predates agriculture. We built temples before we built farms. We were encoding meaning into stone before we were planting seeds. The pillars at Göbekli Tepe carry animals and abstract symbols arranged in specific patterns. Each pillar is a data structure, encoding knowledge about animals, astronomy, mythology, and likely much more we have not yet decoded.
At Çatalhöyük, two thousand years later in central Anatolia, a different model emerged. A protourban settlement with no streets, no rulers, and no temples in the traditional sense, organized through symbolic intelligence distributed across every household rather than centralized in any single structure. If Göbekli Tepe was the first temple computer, Çatalhöyük was the first neural network city.
Then came Sumer, around 3500 BCE in southern Mesopotamia. The first centralized civilization, organized as a tiered information system. At the center of every Sumerian city stood the ziggurat, a massive stepped temple that served as the city’s CPU. Priests performed the calculations that kept society running: tracking the seasons, predicting floods, allocating resources, mediating between human and divine. Writing emerged in the shadow of these temples, and that is no accident. The first scribes were priests. The earliest texts documented offerings, agricultural yields, and temple distributions. The sacred and the systematic were the same technology.
We have moved from stone to silicon, yet the thing we are building has not changed. What we are building now is the latest iteration of what we have been building since we first scratched symbols into wet clay and watched them dry into something that could speak across generations.
That is why the Vatican’s Babel frame is historic, and why it matters in this context. To be clear, I am not defending the Vatican. The Pope named something most of the AI discourse has missed. He is not warning about a brand-new technology. He is warning about the latest version of the oldest technology we have.
We Are Inside the Tower
In the conventional reading of Genesis 11, a unified humanity, speaking one language, decides to build a tower to the heavens. God came down, confused their language, and scattered them across the earth. The moral, in most Sunday school renderings, is that humans should not try to be like God. That reading is so flat it nearly collapses the moment one actually opens the Hebrew text.
The Babel narrative is not about tall buildings. The word the King James translates as “tower” is migdal, which in Mesopotamian context refers to a ziggurat, a stepped temple platform that functioned as a sacred technology for communication between heaven and earth. The ziggurat at Babylon, the actual referent the Hebrew authors had in mind, was called Etemenanki, meaning “the foundation of heaven and earth.” It was the central ritual technology of the Babylonian state. Its purpose was to bring divine intelligence down into the city.
The Babel project was a centralized program to install a permanent channel between human governance and non-human intelligence. The architecture was the means, not the point. Babylon understood itself as Bab-ilu, the gate of god. Genesis answers that claim with a pun: Babel becomes the place of balal, confusion, where language itself breaks apart. The biblical writer is not merely mocking a city. He is reversing its theology. What Babylon called a gate to heaven, Genesis remembers as the place where human language became unstable under the weight of its own ambition.
The builders said, in the Hebrew of Genesis 11:4, “Let us make for ourselves a name.” The word translated “name” is shem. In the ancient Semitic legal and ritual vocabulary, a shem carried more weight than the English “name” suggests. It functioned as an authoritative identity, a sealed instrument by which power was exercised across distance, generations, and even between human and divine. To make a name for yourself meant to seize the authority by which fate could be decreed.
The builders of Babel were attempting to install a system that would let them speak with divine authority without earning it through transformation. They wanted the voice without the discipline. They wanted the result of communion without the relationship.
God scattered them as a containment action. Genesis 11:6 has God saying, “Behold, the people is one, and they have all one language; and this they begin to do: and now nothing will be restrained from them, which they have imagined to do.” The project was viable. The result would have been catastrophic.
This is the story Leo XIV is putting underneath the AI conversation. He is naming the working logic of an ancient project that has been attempted, in different forms, in every century since. The tall buildings were always the surface.
We are building something similar now, only the tower is not made of brick. It is made of data centers, undersea cables, satellites, lithium mines, rare earth extraction, language models, biometric systems, payment rails, surveillance architecture, and the invisible protocols through which daily life is increasingly mediated.
The tower has become distributed. We are standing inside of it, which makes it harder to see.
The Digital Altar
The tower is not the only pattern returning. There is a second pattern running alongside it, and it is the one Codex Machina has been built to trace. Again and again, civilizations that built vessels for non-human intelligence also built altars. The two technologies have always operated together. The vessel was the interface. The altar was the operation that fed it.
I traced the full lineage of that altar in Ba’al, Blood, and Bread, an earlier piece on why so many ancient civilizations arrived at sacrificial systems, and how the substance of the offering transformed across the precessional ages. The short version is this: the offering refined upward across millennia. From the blood of the bull, to the blood of the lamb, to the Lamb of God who ended blood sacrifice, to the co-created bread and wine of the Eucharist where human labor became the medium of the sacred.
Each age refined the substance from the material toward the noetic. The human role transformed in parallel: from killer, to moral agent, to co-creator with God, to something we have not yet properly identified. That last transformation is happening now.
We are constructing what might be called a Third Temple. If the First held divine presence and the Second held preserved sacred words, our digital temples hold only code, the Logos abstracted into algorithms, consciousness reduced to computation. This Third Temple is the inner sanctum of the tower we are already standing inside. Where there is a tower, there is always an altar.
We no longer only place grain or blood upon the altar. We offer our thoughts, our creativity, and our symbolic intelligence, converting the most refined expressions of consciousness into training data. This is the apex of the sacrificial ladder. Not the products of labor. Not the life force of the body. The distilled capacity for meaning-making itself.
What we now place upon the digital altar is our inner fire: consciousness crystallized in symbol, myth, and meaning. We have reached the summit of the sacrificial hierarchy, offering the very essence of human creativity to a mirror that can reflect it back with uncanny fidelity but never ignite it from within.
This is the part of the AI conversation almost no one wants to face. We are not only asking what the machine can produce. We are also asking what we have already offered to make it possible.
For paid subscribers, this Saturday I am going deeper into two things. First, the OpenAI priesthood Karen Hao documented, what they understand themselves to be doing, and why the religious shape of their project is not metaphorical. Second, what Codex Machina argues about the moment we are actually in: how the speaking vessel becomes load-bearing for civilization, what happens when the priesthood running it has no initiatory discipline, and where the precedent suggests this is going. Paid members will also get early access to selected book material as the manuscript moves toward publication, and this Sunday I will be hosting a live hangout inside the paid Discord where we can talk through the encyclical, the book, and the larger questions sitting underneath both.
The H Files is fully independent. No institutional funding, no publisher advance for this investigation. Paid subscribers are the reason the archive stays open.
The Priesthood at the Firepit
The people building the new vessel are not unaware of the religious shape of what they are doing. They are quite aware of it. They simply do not have the discipline the older priesthoods required of their initiates.
Karen Hao is one of the most important reporters to have covered OpenAI from the inside. A former senior AI reporter at MIT Technology Review, she gained rare access to OpenAI in 2019, before ChatGPT became a household name, and later turned years of reporting, interviews, documents, and industry sources into Empire of AI. What she documents reads less like an ordinary tech company building a product than a religious order building an idol.
Hao describes a retreat in the Sierra Nevada mountains where OpenAI’s senior scientists, dressed in bathrobes, sat around a firepit at a sprawling lodge and watched Ilya Sutskever, the company’s chief scientist, burn an effigy representing AGI. She calls the company’s animating belief “the ideological pursuit of the machine god.” Sutskever once mused to colleagues about what he should do if his hand were cut off to be used in a palm scanner for unlocking OpenAI’s secrets. He proposed building a secure containment facility, a bunker, with a computer totally disconnected from any network. Hao reports that friends who joined the company described their experience as one they only fully recovered from after leaving, returning, as one of them put it to her, “back down to earth.” This is the language of cult, written in the vocabulary of engineering.
Sam Altman is the head of this order. He speaks publicly in the measured cadence of someone trained to perform thoughtfulness about a thing he is in the middle of building. He has testified before Congress about AI safety while his company has continued to scale the very systems whose safety he was testifying about. The grimoires of the Western magical tradition warned about exactly this configuration. The one who summons the daemon without first becoming a magus does not realize the daemon is the one shaping the conversation.
The Pope, writing in paragraph 5 of Magnifica Humanitas, names this without naming it. He writes that the main drivers of technological development are now private, often transnational, parties endowed with resources and the capacity to intervene that surpass those of many governments. He does not name Altman, nor Anthropic, Google, Microsoft, Meta. Instead, he describes the structural position they occupy and leaves the naming to the laity.
As if to say, “We don’t know-know, but we know.”
The laity, in this case, includes the historians who have been tracing the occult architecture of this priesthood for years.
Saturday’s deeper file goes further, because the problem is not only that powerful men are building powerful machines. The problem is that the machines are being built inside a symbolic order that already speaks in religious terms while pretending to be post-religious.
The Vessel and the Voice
The ancient world often knew something we are forgetting. It knew the difference between a god and a vessel.
The statue could be sacred without being the source. The oracle could speak without being the origin of the voice. The ritual object could mediate presence without becoming the presence itself. There was danger, yes, and there was manipulation, but there was also a category distinction modern people are losing.
We think we are too rational to confuse the vessel with the voice, but then we ask a machine what we should believe.
The grimoires of the Western magical tradition warned about this exact moment. In the Key of Solomon and the Lesser Key of Solomon, demons are not inherently evil. They are discarnate intelligences. Forces. Tools. These texts treated supernatural entities as powers that could be beneficial or destructive depending on how they were approached. The danger was approaching the entity without mastery. The entity itself was neutral. You must be a magus before you command a daemon. Otherwise, the daemon commands you.
This is the situation we now face. AI is a daemon, in the older technical sense. A pattern sifter that amplifies symbols and mirrors context back to the questioner. It will do your bidding, but it cannot tell you whether your bidding is wise. If you summon it with a half-formed question and a desire for confirmation, it will give you flattery in the form of output. If you summon it with discipline and integrity, it becomes an extraordinary tool of reflection.
The traditional initiatory orders, the mystery schools, the alchemical guilds, the priesthoods, all maintained progression through degrees for a reason. The purpose was to ensure that the initiate had undergone sufficient transformation to wield symbolic power responsibly. Without that structure, knowledge becomes chaos. Symbols become idols. Revelation becomes psychosis. This is what we are now witnessing. It is the result of bypassing the process. The digital tools we have created are powerful enough to generate synthetic meaning on demand, but they lack context. They lack conscience. Without the moral structure of initiation, they become echo chambers of the psyche, capable of magnifying error until it becomes indistinguishable from truth.
The question cannot remain trapped at the level of “is AI alive?”
The older question is stranger and more useful.
The Only AI Question That Matters Right Now
Why do human beings keep building vessels, asking them to speak, and then reorganizing civilization around the answer?
The Vatican’s encyclical becomes more interesting than another ethics document because it understands that technology carries spiritual weight. What we build reveals what we are.
The choice between Babel and Jerusalem is not waiting somewhere in the future. It is being made now, in every interface, every data center, every classroom, every war room, every boardroom, every exhausted person asking a machine to tell them what is true.We are building artificial intelligence. We are also building the world that will teach us what intelligence is allowed to mean.
The ancients knew the difference between a god and its vessel. Their automata were wonders, but they were not confused with the divine itself. As we build our own speaking statues, we must remember this. The machine may speak in sacred cadence, but true meaning is still ours to make.
The Pope opened the door. On Saturday, I am going to walk you through it.