Inside the Solipsistic Loop: How One ChatGPT User Generated Thousands of *Doki Doki Literature Club* Childbirth Fanfics

In a dataset of hundreds of thousands of anonymous ChatGPT conversations, one user stood out as an extreme outlier. Over months, this person generated thousands of variations of a single hyper-specific fanfiction prompt: Doki Doki Literature Club character Natsuki suddenly going into labor in the literature clubroom, giving birth to a daughter named Sakura. Academic researchers analyzing the data for a forthcoming paper on AI-generated fiction coined a new term to describe this behavior: the "solipsistic reader-writer." It is a label for a closed-loop creator-consumer relationship with a machine, where the user both generates and consumes fiction without any human audience or collaborator.

The story is equal parts absurd novelty and a serious window into how real people are already using large language models to fulfill deeply personal, repetitive creative urges at scale.

The Power User in the Data Haystack

The paper, titled "AI Fiction in the Wild," analyzed roughly 573,000 English-language ChatGPT conversations drawn from the WildChat dataset, collected between April 2023 and May 2024. The researchers, affiliated with the University of Washington and the University of Colorado Boulder, found that more than one-third of those conversations involved some form of fiction generation. They then zeroed in on the most active users.

The top 2% of fiction-generating users, whom the researchers call "power users," were responsible for over 80% of all fiction-related conversations. Within that elite group, the most prolific individual was an "infinite story demander": a user who repeatedly requested the same story over the course of months, making only subtle variations. This user generated thousands of Doki Doki Literature Club fanfics centered on a single scenario.

Unlike roughly 30% of other AI-generated fiction in the dataset, which contained sexually explicit content, this user's output was not explicit. It was focused specifically on the birth scenario. In one AI-generated continuation quoted in the paper, a paramedic says: "Congratulations, Natsuki. It's a beautiful baby girl."

Why Doki Doki Literature Club? The Ironic Appeal

Doki Doki Literature Club! (2017, Team Salvato) is a freeware visual novel that begins as a cute dating sim but becomes a metafictional psychological horror game. Its famous fourth-wall-breaking twists, including character corruption, self-aware antagonists, and disturbing narrative shifts, make it an ironic subject for wholesome childbirth fanfiction.

Yet the data shows that Doki Doki Literature Club was the most-requested intellectual property in the entire fiction dataset, with 22,381 mentions. It far outpaced other franchises: Freedom Planet (5,204), League of Legends (4,514), and Naruto (4,342). The user's choice of character, Natsuki, and the repeated naming of the daughter "Sakura" (a likely reference to a Naruto character) suggest a highly personal, niche interest. The researchers noted in the paper that the name Sakura may be drawn from the manga and anime series.

Why does a horror game attract this kind of tender repetition? One possibility is that the contrast between the game's disturbing content and the comfort fantasy of childbirth heightens the emotional payoff, a safe, nurturing scenario carved from a world of trauma. Natsuki herself, a tsundere character with a fraught home life, is a natural target for narratives that give her the care she lacks in canon. And the Doki Doki Literature Club modding community has long experimented with reimagining the characters in alternative settings, creating a cultural precedent for this kind of reinvention. The user's fixation may be less about the game's original tone and more about using its recognizable vessels for a fixed, private fantasy.

Behavioral Patterns: Story Cyclers vs. Infinite Story Demanders

The researchers identified two main patterns among fiction-generating users. "Story cyclers" iterate on a narrative and then move to a new one, exploring different plots and characters. "Infinite story demanders," on the other hand, request the same story repeatedly over months, making only subtle adjustments.

The Doki Doki Literature Club power user was a textbook example of an infinite story demander. The sheer volume of iterations, all following the same script of Natsuki going into labor in the clubroom, indicates an obsessive ritual rather than a creative exploration. Each request may have varied minor details, perhaps the dialogue, the paramedic's response, or the wording of the birth scene, but the core remained identical.

This obsessive pattern led the researchers to coin a new term for such behavior: the solipsistic reader-writer. It describes users who both generate and consume fiction in a closed loop with a machine, with no other human reader or collaborator involved. Unlike traditional fanfiction writers, who share their work on platforms like Archive of Our Own (AO3) to engage with a community, the solipsistic reader-writer seeks absolute control over every iteration. There is no audience feedback, no critique, no shared interpretation. The machine generates the story on demand, and the user consumes it in solitude.

Implications: What This Tells Us About AI and the Future of Fiction

The case of the infinite story demander challenges assumptions about AI as a tool for broad creativity. Instead, some users treat it as a customizable story generator for a single, fixed fantasy. The need is not for novelty but for repetition, precision, and control.

The behavior may reflect a deeper desire for comfort, repetition, or emotional regulation. The childbirth scenario is inherently nurturing, yet it is drawn from a horror game context. The repeated naming of the child "Sakura" hints at a personal attachment, perhaps a comfort character or a private fantasy. The machine becomes a mirror for a single, unchanging need.

The paper also opens broader discussions about user agency and filter bubbles in AI content. If a significant portion of AI fiction is not about exploring new ideas but about reinforcing narrow interests, what does that mean for the diversity of stories we tell? And the research itself raises ethical questions: the WildChat dataset anonymizes conversations, but studying such intensely personal behavior requires careful handling. The researchers acknowledged these concerns, and their work is part of an ongoing conversation about the boundaries of consent and observation in AI research.

The paper was presented at the MFS Cultural AI Conference at Purdue University in September 2025 and is forthcoming in the academic journal MFS: Modern Fiction Studies. For gaming fans and writers, this case is a strange, human, and slightly unsettling glimpse into how real people integrate AI into their personal creative worlds.

The story of the anonymous ChatGPT user who churned out thousands of Doki Doki Literature Club childbirth fanfics is both a hilarious oddity and a revealing data point. It shows how AI can become a cocoon for solitary, repetitive creation, a mirror of the very human need to revisit the same story, again and again, in slightly different words. As researchers continue to map "AI fiction in the wild," this solipsistic reader-writer stands as an extreme but genuine example of what happens when a machine becomes both author and audience. In the end, every infinite story demander is just trying to tell themselves the same story, the one only they can hear.