Superpower Daily

Superpower Daily

Your reading space

Sign in
China Rebukes Anthropic CEO’s Call to Restrict AI Chip SalesOpenAI Says It Will Skip a 2026 IPO Over AI Safety ConcernsOracle Raises Restructuring Estimate to $2.8 Billion as It Cite...MIT Researchers Publish HardFlow to Keep AI Outputs Within Hard...LG AI Research Adds EXAONE Models for Changing Factory WorkUniversity of Washington Finds Fiction in a Third of Sampled Ch...China Rebukes Anthropic CEO’s Call to Restrict AI Chip SalesOpenAI Says It Will Skip a 2026 IPO Over AI Safety ConcernsOracle Raises Restructuring Estimate to $2.8 Billion as It Cite...MIT Researchers Publish HardFlow to Keep AI Outputs Within Hard...LG AI Research Adds EXAONE Models for Changing Factory WorkUniversity of Washington Finds Fiction in a Third of Sampled Ch...
Swipe

University of Washington Finds Fiction in a Third of Sampled Chatbot Chats

The finding points to a committed audience using models less as replacement authors than as responsive engines for role-play and stories tailored to their own choices.

By 2 min read
University of Washington Finds Fiction in a Third of Sampled Chatbot Chats
University of Washington Finds Fiction in a Third of Sampled Chatbot Chats

Listen to this story

The audio brief

About 1:27
0:001:27
Read transcript
More than one-third of the public chatbot conversations sampled in a University of Washington analysis involved fiction generation. That includes prose, poetry, fan fiction, erotica, and role-play—not just attempts to write a conventional novel. Melanie Walsh and Neel Gupta examined more than 500,000 prompts from WildChat, a public dataset of ChatGPT conversations, to look at how people actually use these systems. The striking detail is the concentration. A small group of very heavy users produced a large share of the fictional activity. After accounting for that behavior, the researchers estimated that 7 percent of users engage in fiction generation. That does not mean 7 percent of all chats are fictional, or that those users spend 7 percent of their time on stories. It means custom storytelling is a visible practice among a relatively small, highly active audience. Petra Ferraz de Novaes, a software engineer in Brazil, illustrates what that practice can look like. Using SillyTavern and open-source models, she creates fantasy stories and private role-play scenarios, sometimes reaching the length of a small novel and sometimes moving between original worlds and existing franchises. The product appeal here may be less polished literary output than control: users can change the premise, characters, pace, and next turn. The open question is whether this becomes a distinct entertainment category—and how much the narrow WildChat dataset can really tell us about everyone else.

Story brief

3 key points

Researchers’ strongest finding is not that one-third of all chatbot use is fictional, but that a small, highly active segment drives a large share of the activity. In an analysis of more than 500,000 public WildChat prompts, fiction—including role-play, fan fiction, erotica, prose, and poetry—appeared in over one-third of sampled interactions; adjusting for heavy users reduced the estimate to 7% of users. The result...

  1. 01

    Melanie Walsh and Neel Gupta analyzed public WildChat conversations rather than surveying professional authors.

  2. 02

    The 7% estimate covers users who engage in fiction, not the share of their chats devoted to it.

  3. 03

    Petra Ferraz de Novaes uses SillyTavern and open-source models for novel-length stories and role-play.

For a subset of chatbot users, the appeal is not a polished novel or a one-off answer. It is a story that can change direction on demand, continue indefinitely, and cater to tastes conventional publishing may never serve. A University of Washington analysis of more than 500,000 public chatbot prompts found that more than one-third of sampled interactions involved fiction generation.

The category was broad: prose, poetry, fan fiction, erotica, and role-play all counted. The researchers, Melanie Walsh and Neel Gupta, examined prompts from WildChat, a public dataset of ChatGPT conversations, to understand a behavior often obscured by debate over whether professional authors use AI.

The story is the product

WIRED’s interviews show what that interaction can look like. Petra Ferraz de Novaes, a software engineer in Brazil, uses SillyTavern and open-source models to build fantasy stories and private role-play scenarios. She sometimes produces self-contained stories the length of a small novel; in other sessions, she moves among worlds of her own invention and existing franchises.

A small group writes a great deal

That concentration complicates the headline figure without erasing the pattern. After controlling for heavy use, the team estimated that 7 percent of chatbot users engage in fiction generation. The estimate suggests custom storytelling is not only an extreme hobby, but it does not establish how frequently those users create fiction or what share of their overall chats it represents.

A large signal with a narrow window

That means the research is better read as evidence of a visible, highly engaged practice than as a census of chatbot behavior. Its importance lies in what the practice reveals: some people want fiction not simply to be written for them, but to be continually negotiated with them.

Personal control does not solve the craft problem

For now, that gap helps explain why chatbot fiction may function differently from traditional publishing. The model’s value can be less about literary finish than about responsiveness: it offers a reader control over the premise, pace, characters, and next turn. Whether that leads people toward more reading, away from unfamiliar writing, or simply into a separate entertainment category remains unresolved.

Sources

  1. wired.com‘I Like My Big Rat Wife’: Meet the People Using Chatbots to Write Custom Fiction

Loading discussion...

Related coverage

Continue reading