Generative AI collective behavior needs an interactionist paradigm
Fondazione Bruno Kessler(布雷诺·克塞尔基金会);University of Trento(特伦托大学);Not Diamond;City St. George’s University of London(伦敦圣乔治大学);Carnegie Mellon University(卡内基梅隆大学);Massachusetts Institute of Technology(麻省理工学院);Stanford University(斯坦福大学);Google DeepMind(谷歌DeepMind);University of Chicago(芝加哥大学)
Laura Ferrarotti, Gian Maria Campedelli, Roberto Dessì, Andrea Baronchelli, Giovanni Iacca, Kathleen M. Carley, Alex Pentland, Joel Z. Leibo, James Evans, Bruno Lepri
In this article, we argue that understanding the collective behavior of agents based on large language models (LLMs) is an essential area of inquiry, with important implications in terms of risks and benefits, impacting us as a society at many levels. We claim that the distinctive nature of LLMs--namely, their initialization with extensive pre-trained knowledge and implicit social priors, together with their capability of adaptation through in-context learning--motivates the need for an interactionist paradigm consisting of alternative theoretical foundations, methodologies, and analytical tools, in order to systematically examine how prior knowledge and embedded values interact with social context to shape emergent phenomena in multi-agent generative AI systems. We propose and discuss four directions that we consider crucial for the development and deployment of LLM-based collectives, focusing on theory, methods, and trans-disciplinary dialogue.
OpenHospital: A Thing-in-itself Arena for Evolving and Benchmarking LLM-based Collective Intelligence
School of Software Technology, Zhejiang University(浙江大学软件学院);Laboratory for Statistical Monitoring and Intelligent Governance of Common Prosperity, Zhejiang Gongshang University(浙江工商大学共同富裕统计监测与智能治理实验室);Tongyi Lab, Alibaba Group(阿里集团通义实验室);Alibaba Cloud(阿里云)
GenAINet: Enabling Wireless Collective Intelligence via Knowledge Transfer and Reasoning
Artificial Intelligence and Digital Science Research Center, Technology Innovation Institute(人工智能与数字科学研究中心、技术创新研究所);G Research Center, Khalifa University(6G研究中心、哈利法大学);Centre for Wireless Communications, University of Oulu(无线通信研究中心、奥卢大学);CRAN, Université de Lorraine and CNRS(CRAN、洛林大学和国家科学研究中心)
Towards a new paradigm of scientific discovery with socialized artificial intelligence
Baize Research(白泽研究院);Faculty of Information Engineering and Automation, Kunming University of Science and Technology(昆明理工大学信息工程与自动化学院);Yunnan Key Laboratory of Artificial Intelligence, Kunming University of Science and Technology(昆明理工大学云南省人工智能重点实验室);School of Automation, Southeast University(东南大学自动化学院);School of Artificial Intelligence, Tianjin University(天津大学人工智能学院);School of Computer Science and Technology, Beijing Institute of Technology(北京理工大学计算机学院);SeetaCloud Technology(思特云科技);GPUhub Pte. Ltd.(GPUhub私人有限公司)
Systematic Failures in Collective Reasoning under Distributed Information in Multi-Agent LLMs
School of Computer Science, Carnegie Mellon University, Pittsburgh, USA(计算机科学学院,卡内基梅隆大学,匹兹堡,美国);School of Environment and Society, Institute of Science Tokyo, Tokyo, Japan(环境与社会学院,东京科学研究所,东京,日本)
Indian Institute of Technology Roorkee(印度理工学院鲁尔基分校);New York University(纽约大学);University of Massachusetts Amherst(马萨诸塞大学阿默斯特分校);Georgia Tech(佐治亚理工学院);Princeton University(普林斯顿大学)
Towards Optimal Circuit Generation: Multi-Agent Collaboration Meets Collective Intelligence
National Key Laboratory of Spintronics, Hangzhou International Innovation Institute(国家磁电子键实验室、杭州国际创新研究院;集成电路科学与工程学院,北京航空航天大学);School of Integrated Circuit Science and Engineering, Beihang University(机械电子工程学院,武汉理工大学);School of Mechanical and Electronic Engineering, Wuhan University of Technology(数学与物理科学学院,谢菲尔德大学);School of Mathematical and Physical Sciences, University of Sheffield