Civil Court Simulation with Large Language Models
2026-06-08 • Computation and Language
Computation and Language
AI summaryⓘ
The authors created a computer simulation that mimics how civil court cases are handled, specifically for Chinese law. Unlike criminal cases, civil cases are more complex, so their system uses different roles and steps in the trial process to make it realistic. They found their simulation works well, especially in deciding who is responsible and handling multiple claims. They also showed that keeping good memory of past interactions helps the simulation perform better. Their research explores how different factors like legal rules and social context influence courtroom decisions in the simulation.
civil litigationcourt simulationlarge language modelsmulti-agent systemadjudicationliability allocationstatute retrievaltrial procedurelegal groundingChinese civil law
Authors
Yifan Chen, Haitao Li, Kaiyuan Zhang, Yueyue Wu, Qingyao Ai, Yiqun Liu
Abstract
Court simulation bridges legal education and judicial practice, yet human-based simulations are costly and difficult to scale. Large language models (LLMs) offer a scalable alternative, but existing court-simulation research mainly focuses on criminal cases. Civil litigation is more common in practice and harder to simulate because its claims, liability, and remedies are more flexible. We present a multi-agent court simulation framework for Chinese civil cases. The framework organizes role-based interaction through a five-stage civil trial procedure and integrates memory module and statute retrieval to support long-process adjudication. Experiments show that the framework produces reliable civil judgments, with clear strengths in liability allocation and multi-item adjudication. Further experiments show that memory quality substantially affects downstream simulation quality. Through a five-layer factor framework, we analyze how legal grounding, information conditions, judicial capability and role orientation, organizational pressure, and social context affect the framework's reliability and behavior. These results support the effectiveness of the proposed framework for civil court simulation. The dataset and code are available at: https://github.com/foggpoy/Civil-Court.