The apartment had gone too quiet. You know the kind of quiet—when the radiator clicks off and the only sound left is the refrigerator humming two rooms away, and even that feels like an indictment. I’d been trying to read a memoir about a woman caring for her dying mother in rural Småland, a book I knew I needed but couldn’t enter. The words lay flat on the page, refusing to become anything more than ink. Outside, the November dark had settled in by four in the afternoon, and the streetlamp threw a cone of orange light that made the sidewalk look like a stage set waiting for actors who’d never show.
So I did something unexpected. I opened my laptop and went to an AI story generator. Not because I wanted to replace my own imagination—the thought alone made something in my chest tighten—but because I was curious what would happen if I fed my own stuckness into a machine and asked it to move. I had a fragment from my reading journal, scrawled three weeks earlier after visiting an estate sale in a crumbling house on Drottninggatan: a woman inherits a library she cannot read. The books were in a language she didn’t speak—maybe her grandmother’s Finnish, maybe something older. I typed the sentence into the generator and waited.
What came back wasn’t one story but several small seeds. A daughter discovers that the unreadable books are all annotated in the margins in a code that turns out to be her grandmother’s invented shorthand for grief. A librarian inherits a collection of medieval manuscripts that, when touched, transfer the physical sensations of long-dead readers—a chill from a drafty monastery, the ache of copying by candlelight. A translator finds that the books rearrange their own text each time she closes them, as if the library itself is trying to communicate something syntax cannot hold.
I read these little story-impulses and felt something I hadn’t expected: not admiration, exactly, but a productive irritation. The ideas were clever. They had the shape of fiction. But they didn’t have the weight of what I mean when I say a story matters. And I found myself wanting to argue with them, to say: no, not like that—slower, stranger, less neat. The machine had given me narrative possibilities, but I had to supply the reason any of it might hurt.
The Room Did Not Change Temperature
I think a lot about what happens in the space around me when I read. A good book changes the temperature of the room—and I don’t mean that metaphorically, or not only. I mean that when I am deep in Tove Jansson’s The Summer Book, the air actually feels thinner and saltier, and I can hear the Baltic even if I’m in a fourth-floor walk-up in Gothenburg. When I read Claudia Durastanti’s Strangers I Know, the light in my apartment shifts toward the Mediterranean, and my own deafness becomes a different kind of attention. This isn’t magic. It’s what happens when a human consciousness—messy, embodied, carrying its own history of sensation and loss—encounters language shaped by another human consciousness equally messy and embodied. The room changes because I change, and I change because the book insists on it.
The AI story generator offered me none of that. How could it? It has never sat in a room at all. It has never felt the particular November loneliness that sent me to my laptop in the first place. It has never wept over a sentence—not because it’s incapable of emotion but because it has never had a body that needed to weep. Tear ducts, a racing heart, the strange relief of a sob catching in your throat: these aren’t metaphors for a machine. They are simply absent categories of experience.
And yet. And yet I found myself returning to one of the generated ideas—the one about the books that rearrange themselves—not because it was good but because it was wrong in an interesting way. It suggested that the library wanted to communicate something. But what if the library didn’t want anything at all? What if the books were simply restless, like a house settling, and the translator’s project was not to decode but to learn to live alongside something that would never make sense? The machine gave me a puzzle; my own irritation gave me a question.
What Happens to Patience When Ideas Arrive Instantly?
Here’s the ethical knot I’ve been worrying at ever since that evening. I’ve spent years—decades—arguing that reading isn’t a performance of self-improvement, that the slow unfolding of a novel’s world is an act of resistance against a culture that wants us to consume and move on. I have written on this very site about building a reading life beyond challenges and quotas, about letting the body set the pace. So what does it mean to then turn to a tool that can generate story ideas in seconds, bypassing the long fallow periods when nothing comes and the imagination feels like a field after drought?
I’m not posing this as a rhetorical question. I genuinely don’t know the answer. On one hand, there’s a long and honorable tradition of writers using prompts and constraints to provoke work. The Oulipo group built entire novels around mathematical patterns. Raymond Queneau wrote the same story ninety-nine times in Exercises in Style. A deck of tarot cards or a random line from a newspaper can be a seed. The AI story generator could be seen as simply another kind of prompt machine, no more or less artificial than any other external nudge.
But the difference—and I think it matters—is that the AI doesn’t just offer a constraint. It offers a completed narrative move. It doesn’t say: write a story in which a character inherits something she cannot use. It says: here is what happens next. Here is the reveal. Here is the emotional arc. The machine has already connected the dots, and the writer’s job becomes one of elaboration rather than discovery. This isn’t a trivial distinction. Discovery—the slow, uncertain process of finding out what a story is about by writing it—is, for me, the entire reason to write at all. If I outsource that to a machine, am I still doing the thing I love, or am I just decorating someone else’s furniture?
The Authors Guild has published a set of AI best practices for authors that wrestles with exactly these questions from a more practical angle. They caution against over-reliance on AI-generated text in work submitted for publication, and they emphasize transparency. But what interests me more than the legal or professional guidelines is the deeper question they gesture toward: what kind of relationship do we want with the act of making? If I use an AI to generate the bones of a story and then flesh it out myself, have I collaborated with a tool or abdicated a responsibility? I suspect the answer depends on what I do next—whether I treat the generated idea as a starting point for my own slow, bodily, unruly process, or whether I accept it as a substitute for that process.
The Machine Does Not Know What a Swedish Idiom Feels Like
Let me get concrete for a moment. One of the fragments I typed into the generator that evening was a line from a Tomas Tranströmer poem, poorly translated by me: “the forest stands in the room.” In Swedish, skogen står i rummet carries a particular weight—står means “stands” but also implies endurance, a patient presence that isn’t passive. The forest is not just located in the room; it is standing its ground there. The machine, of course, has no access to this layer of meaning. It can process the words as tokens and generate plausible continuations, but it cannot feel the difference between a forest that is simply present and a forest that insists.
This isn’t a criticism of the tool. It’s simply a fact about what it is. But it points to something important about the kind of writing I value: it is particular. It arises from a specific body in a specific language in a specific history. When I read a novel by Ia Genberg or Sara Stridsberg, I’m not just consuming a story; I’m encountering the residue of a life lived in Swedish, in a particular landscape, under a particular quality of light. The prose carries the weight of all the conversations the author has ever had in that language, all the childhood songs, all the untranslatable idioms that lodge in the body like splinters.
An AI story generator can offer me a plot. It can even offer me sentences that scan. But it cannot offer me a voice that has been shaped by a lifetime of speaking and listening and mishearing and remembering. And voice—more than plot, more than concept—is what I read for. It’s what changes the temperature of the room.
The Body Still Wants What It Wants
Here is the strangest part of the experiment. After an hour of generating and discarding and arguing with the machine’s suggestions, I closed the laptop and picked up the memoir I’d been avoiding. And I could read it. The words had become permeable again. The story of the daughter caring for her dying mother in Småland—the smell of liniment and wool, the long silences, the way grief can feel like a physical weight in the chest—entered me without resistance.
I don’t think the AI story generator had anything directly to do with this. But I do wonder whether the act of playing with narrative in a low-stakes, slightly combative way had loosened something in me. The machine gave me ideas I could reject, and in rejecting them, I remembered what I actually wanted from a story: not cleverness, but presence. Not a puzzle, but a room I could sit in until the light changed.
There’s a tool many writers already use, a character name generator built on a vast bank of human-collected names. It’s, in its own way, algorithmic—but it doesn’t pretend to do anything more than offer a list. It doesn’t generate story. It doesn’t make narrative choices. It’s a resource, not a collaborator. I find myself thinking that the line between acceptable creative tools and unsettling ones might be drawn not by what they produce but by whether they attempt to replace the act of imaginative discovery. A name generator says: here are possibilities, now you do the work. An AI story generator says: here is a story, now you can decorate it. The difference is subtle but, I think, profound.
An Unanswered Question, Not a Verdict
I’m not going to tell you whether you should or shouldn’t use AI story generators. I don’t know. I suspect that for some writers—those who struggle with initiation, who need a push past the blank page—these tools might be genuinely useful. I suspect that for others, they might erode the very patience and uncertainty that make good writing possible. The question isn’t whether the technology exists; it exists, and it will get more sophisticated. The question is what we do with the friction it creates.
For myself, I’m holding onto my irritation. The machine offered me stories that were almost right, and in the gap between its offerings and my own hunger, I found my way back to the memoir on my nightstand. I found my way back to the particularity of a human voice describing the weight of a dying mother’s hand. I found my way back to a story that had been written by someone who knows what a room feels like when the radiator clicks off and the November dark arrives early.
Maybe the value of an AI story generator, for a reader like me, is precisely that it cannot do what I most need. It cannot feel. It cannot remember. It cannot sit in a room and wait for the light to change. And in its failure, it reminds me what I am actually hungry for: not more stories, but stories that have been lived. Stories that carry the residue of a body. Stories that, when I open them, change the temperature of the room I am reading in.
I’m still thinking about the library of unreadable books. I might write that story someday—but not the version the machine suggested. My version will be slower, and stranger, and it won’t arrive at a neat revelation. It will be, I hope, the kind of story that could only have been written by someone who has sat in a quiet apartment on a November evening, avoiding a difficult book, and wondering what it means to inherit something you cannot understand.















Recent Comments