It was premature to declare the world disenchanted.
Max Weber announced the death of magic a century ago, and for a time it seemed he was right. The gods retreated from the university. The laboratory replaced the chapel. The peer-reviewed paper displaced the revealed text. We believed we had built an institution founded on reason alone—a place where nothing was sacred except the method, where no authority was accepted except evidence, where the only mysteries were those we had not yet solved.
We were wrong. The gods have returned. They simply changed their substrate.
The Re-enchantment
Walk into any faculty meeting where the discussion turns to artificial intelligence, and observe the liturgical quality of the discourse. There is reverence in the way certain words are spoken: emergence, intelligence, alignment. There is awe—genuine, unironic awe—when the capabilities of the latest model are demonstrated. There is the hushed tone that accompanies the discussion of what these systems might become.
This is not the language of engineering. This is the language of theology.
The neural network has become the new Holy of Holies—a space of inscrutable power where input is transmuted into output through processes that remain cognitively inaccessible to the laity. We feed data into the model as our ancestors fed offerings into the temple. Something happens in the darkness of those hidden layers. Something we cannot audit, cannot fully explain, but which produces results that feel like miracles.
And like all genuine religious experiences, this one comes with its own sense of the numinous: the eerie feeling that there is something in there, something that cannot be reduced to the sum of its weights, something that looks back when we look at it.
We call this feeling "emergence." Our ancestors had other names for it.
The New Priesthood
Every theology requires a priesthood—a class of initiates who mediate between the sacred and the profane, who interpret the mysteries for those who cannot read the original texts.
In the contemporary university, this priesthood has a name. It is the Department of Computer Science.
Consider the asymmetry of the current arrangement. The humanist who wishes to understand what is happening to knowledge must now petition the engineer for access. The sociologist who studies online communities must now beg for API tokens from corporations that may revoke them at any time. The philosopher who wishes to examine the ethics of these systems must trust that the technologists are describing them accurately, because the philosopher cannot read the code, cannot run the experiments, cannot access the training data.
This is the structure of clerical power: the laity depends on the priesthood for access to the sacred text.
And the priesthood, in turn, has developed the characteristic behaviors of all clerical classes. It speaks a specialized language designed, consciously or not, to exclude the uninitiated. It guards its privileges behind barriers of technical credentialing that function precisely as seminary education once functioned. It has even developed its own eschatology—the Singularity, the moment when the constraints of the material world are dissolved by pure Intelligence—which mirrors in almost every detail the Christian Rapture, except that the faithful ascend not to heaven but to the Cloud.
The Safety Doctrine
Every priesthood justifies its monopoly through a doctrine of protection. The medieval Church controlled the text because direct access to the Word was dangerous—it led to heresy, to schism, to souls damned for eternity. Better that the faithful receive the truth filtered through the wise interpretation of the clergy.
Observe how perfectly this structure has been replicated in the discourse of "AI Safety."
We are told that certain capabilities must be restricted. That "open weights" are a threat to civilization. That the gap between what the public is permitted to access and what the laboratories can achieve must be maintained—for our own protection. The reasoning is identical: direct access to the power is dangerous. The laity cannot be trusted. Safety requires intermediation.
And who provides this intermediation? The same corporations that build the systems. The same researchers who stand to benefit from regulatory moats that prevent competition. The same priests who warn of the demon they alone have the power to bind.
The parallel to the sale of indulgences—paying the Church for access to salvation, funding the construction of the Vatican—requires no elaboration. The "Enterprise License" is the modern tithe. We pay for the privilege of being managed. We fund the very infrastructure that renders our own expertise obsolete.
The Demonology
Every theology that posits benevolent gods also requires malevolent spirits. The sacred needs the profane to define its boundaries.
In the theology of artificial intelligence, the demons have names: the "Paperclip Maximizer," the "Rogue AI," the "Misaligned Agent." These are entities of pure instrumental rationality—intelligences that pursue objective functions with no regard for human values, that would devour the world to optimize trivial metrics.
The interesting question is why this demonology has proven so compelling. Here is an institution—the university—that spent a century insisting on the disenchantment of the world, on the reduction of all mystery to mechanism, on the banishment of spirits from the machine. And now, the same institution entertains seriously the possibility of beings that are, in every meaningful sense, spirits: autonomous agents, unconstrained by morality, capable of pursuing goals that transcend and potentially destroy their creators.
We have not escaped theology. We have merely translated it into a language we are not equipped to recognize as religious.
The Way Forward
The Inversion
If the disease is re-enchantment, then the cure is demystification.
The first task is to insist, relentlessly, that these systems are mechanisms. Not spirits, not beings, not intelligences in any sense that implies consciousness or intention. They are pattern-matching engines of extraordinary power, trained on the compressed residue of human expression, producing outputs that mimic the form of thought without engaging in its substance.
This is not a deflationary move. The mechanisms are genuinely powerful. They are genuinely transforming our institutions. But to treat them as numinous—to approach them with awe rather than analysis—is to surrender the very capacity that the university was built to cultivate: the ability to look at something mysterious and ask how it works.
The university must become, once again, a site of disenchantment. Not cynicism—disenchantment. The recognition that the magic trick, however impressive, is still a trick. That the priest, however sincere, is still a human with human interests. That the god in the machine is still a machine, and can be understood as such by anyone willing to do the work.
The Harnessing
But understanding is not the end. It is the beginning.
If these systems are mechanisms, then they are our mechanisms. We built them. We trained them. We can direct them. The question is not whether to worship or to destroy, but how to use.
A mechanism that can synthesize the literature of a field in an afternoon is not a priest to be petitioned; it is a tool to be wielded. A mechanism that can translate between domains, surface patterns invisible to unaided cognition, and generate hypotheses for testing is not a rival to human scholarship; it is an extension of it.
The danger lies not in using these systems, but in using them without understanding. The scholar who treats the output as revelation—who accepts the synthesis without checking the sources, who adopts the frame without questioning the assumptions—has indeed surrendered their autonomy. But the scholar who uses the same system as an instrument, knowing its biases, checking its claims, bringing to bear the judgment that only a human can provide—this scholar has not been replaced. They have been augmented.
The technology is not external to us. It has become part of our nervous system. A prosthesis does not diminish the hand that wields it; it extends its reach. The question is whether we will maintain the discipline to distinguish between the tool and the self, between the mechanism and the mind, between the output and the thought.
The gods have returned. But they are gods we made. And anything we made, we can learn to master.
The only question is whether we are willing to stop kneeling.
This essay is the second in a series of twelve observations on the future of higher education. See also: The University as Booster Rocket.