Time to Talk: LLM Agents for Asynchronous Group Communication in Mafia Games

TL;DR

This paper introduces an adaptive asynchronous LLM agent designed for group communication in Mafia games, demonstrating its ability to perform comparably to humans and mimic human-like speaking patterns in a social game setting.

Contribution

The work presents a novel asynchronous LLM agent with separate generation and scheduling modules, evaluated on real-world Mafia game data, advancing LLM application in social and group contexts.

Findings

Agent performs on par with humans in game metrics

Agent's speaking timing closely mirrors human patterns

Differences observed in message content compared to humans

Abstract

LLMs are used predominantly in synchronous communication, where a human user and a model communicate in alternating turns. In contrast, many real-world settings are asynchronous. For example, in group chats, online team meetings, or social games, there is no inherent notion of turns. In this work, we develop an adaptive asynchronous LLM agent consisting of two modules: a generator that decides what to say, and a scheduler that decides when to say it. To evaluate our agent, we collect a unique dataset of online Mafia games, where our agent plays with human participants. Overall, our agent performs on par with human players, both in game performance metrics and in its ability to blend in with the other human players. Our analysis shows that the agent's behavior in deciding when to speak closely mirrors human patterns, although differences emerge in message content. We make all of our code and data publicly available. This work paves the way for integration of LLMs into realistic human group settings, from assistance in team discussions to educational and professional environments where complex social dynamics must be navigated.