Enhancing Consistency of Werewolf AI through Dialogue Summarization and Persona Information

2026-03-07

AIWOLFDIAL

AI Summary

Existing AI agents for Werewolf often exhibit inconsistent statements and role-character drift across multiple rounds of dialogue, undermining their reasoning credibility and gameplay performance. This work proposes a novel approach that integrates large language model–generated dialogue summaries with handcrafted role profiles—including linguistic style templates and example utterances—to explicitly incorporate role-specific information into a contextual summarization mechanism. By fusing explicit role identity with dynamic context tracking, the method enhances long-term behavioral and linguistic consistency throughout the game. Experimental results demonstrate that the resulting agents significantly improve contextual coherence and role stability in self-play settings, effectively preserving consistent speech patterns and strategic behaviors aligned with their assigned roles.

Technology Category

• Multiagent Systems: Adversarial Agents
• Humans and AI: Game Design — Virtual Humans, NPCs and Autonomous Characters
• Natural Language Processing: Conversational AI/Dialog Systems

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI

Abstract

The Werewolf Game is a communication game where players’ reasoning and discussion skills are essential. In this study, we present a Werewolf AI agent developed for the AIWolfDial 2024 shared task, co-hosted with the 17th INLG. In recent years, large language models like ChatGPT have garnered attention for their exceptional response generation and reasoning capabilities. We thus develop the LLM-based agents for the Werewolf Game. This study aims to enhance the consistency of the agent’s utterances by utilizing dialogue summaries generated by LLMs and manually designed personas and utterance examples. By analyzing self-match game logs, we demonstrate that the agent’s utterances are contextually consistent and that the character, including tone, is maintained throughout the game.

Problem

• Werewolf AI
• consistency
• dialogue summarization
• persona information
• utterance coherence

Innovation

• dialogue summarization
• persona information
• Werewolf AI
• large language models
• consistency enhancement