LLM Court is a serious game designed to simulate a courtroom environment using agents powered by Large Language Models (LLMs). The project aims to provide a training tool for students who wish to practice trial dynamics, and offering legal professionals an interactive setting to refine their argumentative skills and defensive strategies. The system integrates Large Language Models within a gameplay structure inspired by courtroom dramas, combining elements such as witness examination, turn management, testimony analysis, and verdict generation by the model.
LLM Court è un serious game progettato per simulare un’aula di tribunale sfruttando agenti basati su Large Language Models (LLM). Il progetto mira ad offrire uno strumento formativo per studenti che desiderano esercitarsi con le dinamiche processuali e, al tempo stesso, fornire ai professionisti del settore legale un ambiente interattivo per affinare capacità argomentative e strategie difensive. Il lavoro integra Large Language Models all’interno di una struttura di gioco ispirata ai courtroom drama, combinando elementi di interrogazione, gestione dei turni, analisi di testimonianze e valutazione tramite verdetti generati dal modello.
LLM Court: Progettazione, sviluppo e valutazione di un serious game per la simulazione forense basato su agenti AI
BOLOGNINI, MASSIMILIANO
2024/2025
Abstract
LLM Court is a serious game designed to simulate a courtroom environment using agents powered by Large Language Models (LLMs). The project aims to provide a training tool for students who wish to practice trial dynamics, and offering legal professionals an interactive setting to refine their argumentative skills and defensive strategies. The system integrates Large Language Models within a gameplay structure inspired by courtroom dramas, combining elements such as witness examination, turn management, testimony analysis, and verdict generation by the model.| File | Dimensione | Formato | |
|---|---|---|---|
|
LLM_Court.pdf
embargo fino al 15/06/2027
Dimensione
6.32 MB
Formato
Adobe PDF
|
6.32 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.12075/24678