< Explain other AI papers

BookWorld: From Novels to Interactive Agent Societies for Creative Story Generation

Yiting Ran, Xintao Wang, Tian Qiu, Jiaqing Liang, Yanghua Xiao, Deqing Yang

2025-04-23

BookWorld: From Novels to Interactive Agent Societies for Creative Story
  Generation

Summary

This paper talks about BookWorld, a new system that takes information from novels and uses it to create virtual societies filled with different characters, allowing for the creation of new, creative stories that feel true to the original books.

What's the problem?

The problem is that most AI story generators either make up random stories that don’t really match the style or logic of real books, or they can’t handle complex interactions between lots of characters like you’d find in a real novel.

What's the solution?

The researchers built BookWorld, which can read and understand novels, then use what it learns to set up a world filled with agents (like characters) who interact with each other. This setup lets the AI generate stories that are both creative and faithful to the original material, and it can also be used to simulate social situations or test out different story ideas.

Why it matters?

This matters because it can help writers, game designers, and researchers create richer, more believable stories and virtual worlds, making creative projects more interesting and realistic.

Abstract

BookWorld, a system for simulating multi-agent societies based on books, generates high-quality, faithful stories and supports applications like story generation and social simulation.