EB-MAS：为 LLM 多智能体系统补上可审计信念更新的安全层
Auditable Belief Updates as a Missing Safety Layer for LLM-Powered Multi-Agent Systems
AI 导读
arXiv 论文提出显式信念多智能体系统（EB-MAS），将 LLM 智能体的信念状态外化为结构化载体，用信念算子提出候选更新，并通过 Cognitive Firewall 拦截证据不足却受脆弱性线索驱动的更新。论文还讨论群体信念风险，如制造共识、共谋与错误信息级联，指出智能体安全需在运行时对信念修正进行追踪、校准和约束。
正文 · 原文
Abstract:Large Language Models (LLMs) are increasingly being used as cognitive interfaces that shape how users interpret evidence, communicate, and make decisions. As deployed systems become more multimodal, tool-using, and multi-agent, safety failures no longer arise only from isolated harmful outputs, but also from how agents update, store, and propagate epistemic commitments across retrieval, memory, tools, and peer communication. In this paper, we define belief as an engineering abstraction for evidence-based commitments that guide downstream action. Building on this view, we propose the Explicit Belief Multi-Agent System (EB-MAS), which externalizes belief states into a structured substrate, uses belief operators to propose candidate updates, and applies a Cognitive Firewall to bound or reject updates that are weakly evidenced but strongly driven by vulnerability cues. We also discuss the risk of group belief at the group level, including manufactured consensus, collusion, and cascading misinformation, and outline practical integration challenges around grounding, false positives, scalability, and ethical misuse. Therefore, scalable agent safety requires not only better output filtering or stronger base models, but also explicit interfaces for tracing, calibrating, and constraining belief revision at runtime.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.38869 [cs.AI] |
| (or arXiv:2610.38869v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.38869 arXiv-issued DOI via DataCite |
Submission history
From: Zhuoyun Li [view email]
[v1]
Wed, 28 Oct 2026 20:40:33 UTC (1,210 KB)
来源：arXiv cs.AI · arxiv.org