摘要
生成式人工智能技术的突破性发展为社会模拟方法的革新带来了新机遇。本文从社会模拟方法的基本要素出发,探究生成式人工智能如何凭借其在自然语言理解、多模态生成和情境适应等方面的关键特性,为智能体建模范式的拓展与升级创造新的可能。在系统梳理现有研究成果的基础上,本文围绕典型任务类型与应用场景,对“生成式智能体建模”的实践路径与建模模式进行了分类、归纳与讨论。同时,从社会科学的研究视角出发,进一步探讨生成式人工智能融入社会模拟过程中可能存在的理论挑战与潜在风险,并对未来发展方向进行展望。本文旨在系统梳理生成式人工智能赋能社会模拟研究的方法,以推动社会科学研究范式的融合与创新。
The breakthrough development of generative artificial intelligence has brought new opportunities for the innovation of social simulation methods.Starting from the fundamental elements of social simulation,this paper explores how generative AI,through its key capabilities in natural language understanding,multimodal generation,and contextual adaptation,opens up new possibilities for expanding and upgrading agent-based modeling paradigms.Based on a systematic review of existing research,the paper categorizes and discusses practical pathways and modeling patterns of“generative agent-based modeling”across typical task types and application scenarios.Furthermore,from the perspective of social science research,the paper examines theoretical challenges and potential risks that may arise when integrating generative AI into social simulation.It concludes with an outlook on future development directions.This paper aims to provide a systematic overview of how generative AI empowers social simulation research,in order to promote the integration and innovation of paradigms in the social sciences.
作者
王志超
吕泽宇
Wang Zhichao;Lyu Zeyu
关键词
生成式人工智能
基于智能体建模
大语言模型
社会模拟
generative artificial intelligence
agent-based modeling
large language models
social simulation
