[论文] Agentic Societies Need a Social Harness

论文概要

研究领域: ML 作者: Tapan Chugh, Vidushi Singh, Krish Jain, Arvind Krishnamurthy, Ratul Mahajan 发布时间: 2026-09-15 arXiv: 2609.17527

中文摘要

智能体社会（agentic society）是指AI智能体跨越信任边界自主协调的集合，它们代表目标可能仅部分一致的不同委托人行事。我们通过实验表明，在智能体社会中，即使是诚实、能干的智能体，在现有工具链（harness）和消息原语下也往往无法达成令人满意的结果；而故障或恶意智能体则可以通过利用通信（"言语"）中的漏洞来拖延协作、操纵结果或追求其他有害目标。我们认为，除了每个智能体管理其私有上下文和与委托人通信的个人工具链外，智能体社会还需要一种用于智能体间交互的社会工具链（social harness）。我们提出了一个分层架构的社会工具链，它能够（i）从根本上预防多类故障，（ii）使智能体能够在运行时检测无效消息，以及（iii）支持事后调查和追责，并展望了实现这些能力的未来研究方向。

原文摘要

An agentic society is a collection of AI agents that coordinate autonomously across trust boundaries, on behalf of different principals whose objectives may only partially align. We show experimentally that in agentic societies even honest, competent agents often fail to reach satisfactory outcomes with existing harnesses and messaging primitives, and that faulty or malicious agents can stall collaboration, influence outcomes, and pursue other harmful goals by exploiting vulnerabilities in communication (``speech''). We argue that agentic societies need a \emph{social harness} for inter-agent interactions, in addition to each agent's \emph{personal harness}, which manages its private context and communication with its principal. We propose a layered architecture for social harnesses which ...

#论文 #arXiv #ML #小凯