# AI and the Ritual Society AI makes the rituals of reports, creative work, and retrospectives cheap and polished. In doing so, it forces us to ask where expertise, understanding, and judgment actually reside. ## Metadata - HTML: https://glenzli.com/en/notes/ai-and-ritual-society/ - Markdown: https://glenzli.com/en/notes/ai-and-ritual-society.md - Collection: Notes - Language: en - Published: 2026-07-02 - Updated: 2026-08-17 - Tags: ai, society, judgment, workflow ## Content AI is changing many things. Whether it will make the world better remains to be seen. But the world has certainly become much better at looking “good.” Once, even producing something that merely looked “good” took serious effort, whether it was a report, a research project, a work of art, or a performance. Now all it takes is a wish. A model can produce these “good” things in bulk: polished formatting, professional structure, refined details—often more convincing than what most people could produce on their own. So a question begins to surface: > When doing something “good” becomes this easy, what did we value in “good” in the first place? AI did not create this problem. It merely dragged our hidden pretenses into the open. ## Ritual and Reality Human attention and energy are painfully finite. Over thousands of years, we built processes, standards, techniques, and formats around that limit. These structures have helped us produce excellent work. They are not bad things. Often they are indispensable. Processes and standards help people collaborate, check results, and take responsibility. Techniques and formats help work be presented, understood, and shared. The trouble begins when these motions drift away from the ends they were meant to serve. There is no real understanding, but the study notes are immaculate. The problem remains unsolved, but the retrospective is profound. The research has stalled, but the format and technique are flawless. Risk has not gone down, but the review is complete. At that point, these structures stop helping good work happen and start serving as proof that it already has. We stop asking: > Did anything actually change? It becomes easier to ask: > Was every part of the ritual completed? ## Barriers Made of Ritual Many barriers to entry have long been less substantial than they look. On paper, entering a field requires knowledge, experience, and professional ability. In practice, much of the barrier consists of expression, format, jargon, procedure, and posture. You need to know how to write the paper, arrange the proposal, present the report, and anticipate what reviewers want. You also need the local vocabulary: “research gap,” “risk closure,” “reusable methodology,” “long-term value.” None of this is meaningless. A field needs its own standards and structures. Complexity itself is a defense: it keeps the field from being instantly flooded by low-effort imitation and spam. But once the rituals become complicated enough, fluency in ritual is easily mistaken for mastery of the field. Worse, in some fields ritual is no longer just an outer shell. People spend enormous effort learning how to make something look right while asking less and less about the thing itself. Real experience can disappear when it arrives in the wrong format. Empty language passes easily when it arrives in the right one. At that point, ritual begins to replace real ability. It becomes a stand-in for the field itself. For a while, reality and ritual held a kind of balance. Reality was not entirely buried, while ritual did enough to protect and sustain the field. Then AI arrived. Balance gone. AI is too good at ritual. Papers, reports, proposals, works of art; professional or amateur, scientific or mystical—it can give all of them a convincing surface. Ritual went from barrier to speedrun. Showing up with AI does not mean someone has mastered the field. But once the ritual barrier collapses, the question underneath becomes painfully visible, enough to unsettle many systems: > Where does a field’s real expertise actually live? Clearly, not in ritual. Yet some fields have reached a point where, once the ritual is peeled away, there is nothing inside. ## From Learning to Life The erosion does not stop at professional fields. Take an extreme example. At a lecture, one person fills page after page with notes and still cannot answer a basic question. Another writes nothing down and leaves with a coherent understanding. The notes are not necessarily a performance. The learning simply did not happen in them. From link aggregators and RSS feeds to today’s AI knowledge bases, we have kept repeating one ritual of learning: if I collect enough information, I will become smarter. AI performs this ritual beautifully. A concept has never entered anyone’s mind, yet the notes and study plan are already complete. A problem remains unsolved, while its retrospective and process record keep growing. The material accumulates. Judgment may not. A new kind of illiteracy follows. Someone may recognize every word and save every article and document, yet be unable to make sense of a sentence without an AI summary. Even after the summary, what remains may be nothing more than inert text, immediately fed forward as the next prompt to another AI. Today, ideas need not collide in a human mind, and knowledge need not be worked through there. But they almost certainly pass through AI. An idea can be made to look ever more like a paper without ever being tested or discussed. A dazzling work can appear from nowhere without a human spark behind it. Academia is no exception. Neither is life. AI can help plan travel, diet, exercise, and a career. The convenience is real. But a person may spend less and less time facing the question “What do I actually want?” and more time carrying out a life plan that merely looks reasonable. Not all of this is bad. AI makes it easier to begin and lets people give concrete form to impulses that used to remain vague. But form now comes too easily. Keep pressing Continue, and a polished result you once could not have made even after giving it everything you had appears on the screen. It is hard not to claim the result as your own, even if it never existed in your mind. In the end, learning and life may be reduced to: > AI summarized it. AI made the plan. ## The Evolution of Ritual Society tends to reward visible traces. Judgment, understanding, and whether anything has truly moved forward are hard to see. Reports, meetings, process records, and slide decks are easy to count. Everyday life works the same way: how a person is actually living is hard to measure, while check-ins, photos, and posts are right there on the screen. Once the trace becomes the target, people optimize the trace. That is Goodhart’s law. It is an old story. Given enough time, a ritual society takes shape almost by itself. Many people once objected to ritualization because it stood in the way of the real work. Faced with the machinery of society, they bent to it anyway. Then AI arrived and the friction of ritual suddenly disappeared. Many former rebels turned around and eagerly joined in building rituals at a much larger scale. That raises an uncomfortable possibility: some resistance may have been less a rejection of ritual than an inability to perform it. AI makes the traces left by ritual cheaper, prettier, and harder to distinguish from real progress. Once everyone can perform the ritual with ease, the reason to stand outside it disappears as well. The result is quiet and dangerous: > The real problem remains unsolved, while the ritual around it grows ever more complete. Once beautiful rituals become commonplace, they no longer look beautiful. Naturally, the next move is to draw another magic circle and fire up the furnace: AI reads the material, distills experience, reflects on failure, proposes a new strategy, and repeats the cycle in pursuit of self-evolution. Reading material, generating plans, comparing paths, and proposing candidates are exactly the kinds of work AI does well. But without a real task, clear acceptance criteria, a record of failure, and a cost limit, this “evolution” easily becomes a ritual loop. Failure produces a retrospective. The retrospective produces a new plan. When the new plan fails, an even better retrospective follows. The magic circle can evolve nothing but purer ritual. The outcome that ritual once served no longer participates in selection. Darwin’s coffin lid must be nailed shut before evolution can be allowed to run this wild. ## The Judgment We Give Away AI’s greatest temptation may not be to create for us, but to make us willingly surrender judgment. Judgment is exhausting. It requires us to rank things, make tradeoffs, and absorb a great deal of information. It also requires us to admit that we may be wrong, and that some failures contain no lesson worth preserving. Sometimes what people want to automate is not the details of the work, but judgment about the work itself. Let AI define the goal, plan the path, verify the result, and carry the complexity; the person keeps the payoff, the status, and the right to sign off. But if a system can independently define goals, judge value, choose paths, accept results, and take responsibility, what is the person who contributes only a signature still doing in the system? Once judgment has been handed away, retaining the signature does not mean retaining control. That may be more troubling than the familiar question of whether AI will replace people. There is another difficulty: on many everyday tasks, AI’s answer is already more reliable than the snap judgments many people make. Choosing to preserve your own judgment means accepting the fatigue of thinking, the embarrassment of being wrong, and sometimes doing worse than AI. It is not a comfortable choice. ## The Ritual Engine The AI era will not automatically make society smarter. It may first make society better at looking smart. More reports, fuller processes, prettier retrospectives, more impressive work—none of it necessarily means more judgment. AI itself is not the culprit. For someone with a clear goal and a willingness to verify the result, it is a useful tool. But in a society already inclined to reward ritual, it will amplify that inclination. So perhaps the most important question in the AI era is not: > Can we get AI to do more? It is: > Do we still know what is worth doing? If we do, AI is a tool. If we do not, it will help us perform the ritual more beautifully than ever.