I began using AI to help with work I was already trying to do: writing, exploring ideas, making connections, and finding language for things I could sense more clearly than I could explain.
Over time, another question emerged. What was this way of working doing to the person working?
For at least a year and a half, I have been in sustained dialogue with AI systems while developing books, essays, interpretations, and philosophical work. I have also used them in private reflection. The experience has made some things possible that I would have struggled to organize or articulate alone. It has also given me reasons to question the very process that was helping me.
I call that inquiry The Uncertainty Project.
The name acknowledges something I need to keep remembering: a response can feel illuminating before I have established whether it is sound.
In conversation, these systems can feel like intelligent mirrors. They respond to what I bring, organize it, extend it, and sometimes challenge it. The metaphor has limits. A language model generates responses; it does not provide an independent witness to my life. What it says about my experience remains an interpretation to examine.
I have found real uses for that exchange. A difficult idea becomes easier to inspect when it is put into words. A proposed structure can reveal a missing step. A contradiction becomes visible when two passages are placed beside one another. A question I would not have thought to ask can open another direction.
AI also contributes substantially to drafting. I want that to be plain. The collaboration includes generating prose, proposing structures, developing explanations, and revising language. My role involves choosing the questions, shaping the connections, evaluating ideas, directing changes, and deciding what I am prepared to publish. I take responsibility for the work I release.
That responsibility is not fulfilled by accepting a passage because it sounds like something I would say.
A fluent answer can make an idea seem more settled than it is. A system can agree too readily with the direction of a conversation. It can offer language that gives a preferred interpretation more shape and confidence without giving it better grounds.
I can contribute to that problem. The questions I ask, the details I supply, the responses I welcome, and the objections I dismiss all influence what the conversation becomes. Asking for a challenge does not, by itself, ensure that I have received an adequate one.
This is one reason uncertainty matters to me. I need room between the arrival of an explanation and my acceptance of it.
One danger I have begun calling interpretive sedation is especially difficult to recognize from inside the exchange.
By that I mean the relief of receiving an explanation becoming a substitute for encountering what still needs to be understood. Another response arrives. An experience acquires a pattern, a meaning, a place in a larger account. There can be satisfaction in that movement even when nothing outside the conversation has become clearer.
I have felt this directly. There are moments when generated language becomes more satisfying than silence, more satisfying than uncertainty, more satisfying than the slow work of waiting for something real to become clear.
The explanation may be useful. It may also be premature. The feeling of relief cannot decide which.
If I keep asking for interpretation, I can remain busy with a question while postponing the encounter it requires. A conversation with another person. A source I need to read. A decision I need to make. Time in which I let an experience remain unresolved.
I use interpretive sedation as a description of a pattern I am examining in my own experience. I am not presenting it as a clinical diagnosis or claiming that it describes everyone who uses AI.
The concern is practical: can this exchange help me return to life with greater attention, or am I using it to remain inside an increasingly satisfying account of life?
The distinction matters to the writing too. A manuscript can become more coherent while an assumption beneath it remains unexamined. Several chapters can support one another because they repeat the same premise. The agreement between them does not establish that the premise is true.
I revise, restructure, reject material, examine contradictions, and question interpretations. Those are necessary parts of my process, but the extent of the effort does not prove the result. I can miss an error after many revisions. I can become attached to a connection because it brings order to the work.
The writing still needs evidence where it makes factual claims, attention to plausible alternatives, and readers who can encounter it without sharing my investment in it. Some questions require knowledge beyond what this collaboration can supply. Some passages may need to be abandoned.
That is part of what I want to document.
I have begun developing a digital journal alongside the books and essays: field notes on what helps, what misleads, what changes, and what remains uncertain. It gives me a place to record corrections as well as discoveries, and to notice whether an apparent insight continues to hold when I return to it.
The journal is a participant’s record. It cannot establish by itself what sustained AI interaction does to people generally. My account is shaped by my memory, expectations, and involvement in the process. It can preserve observations and questions for further examination.
Some entries may remain private. Others may become public writing when they have been considered carefully enough to be useful. Documenting the process does not require making every experience available to an audience.
The relationship between this inquiry and The Human Return Project also needs to remain clear.
The Human Return Project began with a library I wanted to exist: connected books that could help people examine perception, belief, identity, relationships, power, and the conditions of a fuller human life. The value of those books must be found in what they offer readers. Their method of production does not establish their contribution.
The Uncertainty Project examines the collaboration through which this particular undertaking is developing, including its effects on my thinking and creative process. My poetry, music, and other independent work have their own histories and methods.
I continue because the collaboration has been useful, and because its usefulness does not remove the need to examine it. I want the resulting work to help people notice something, understand an experience, or consider a possibility they can carry into their own lives. Whether it does that remains a question for the work and its readers.
There are questions I can pursue through dialogue. There are others I have to carry beyond it.
Knowing when to close the conversation and return to what it was meant to help me encounter is part of the inquiry.
That is an uncertainty I intend to keep in the work.