A study analyzing user interaction data reveals that a substantial portion of consumer chatbot activity is dedicated to creative writing, fan fiction, and personal role-play. Researchers at the University of Washington examined over 500,000 anonymized ChatGPT prompts from the public WildChat dataset and found that over one-third of analyzed interactions involved generating or refining fictional narratives.
While high-volume power users accounted for a significant share of the total fiction prompts, researchers estimated that approximately 7 percent of overall chatbot users engage in fiction generation. Users frequently leverage open-source models and front-end tools like SillyTavern to generate ongoing, interactive storylines across custom and existing fictional universes.
These findings highlight a contrast between public pushback against commercial AI-generated books and the quiet adoption of LLMs as interactive storytelling companions by individual consumers.
Why it matters
Highlights persistent consumer demand for interactive, customized LLM narrative applications and role-play platforms.
Validates open-source model deployment for niche consumer software interfaces and localized text generation tools.
Provides empirical usage metrics counter to broader media narratives surrounding consumer rejection of AI-generated prose.
Source: wired.com
