Risks and Controls for Multi-Agent Systems: an analytical framework for deployment of AI agents across organisational boundaries
多智能体系统的风险与管控:跨组织边界部署AI智能体的分析框架
专题命中 多智能体 :agent(title,abstract);AI agent(title,abstract);multi-agent(title,comments);分类 cs.AI
AI总结 本研究提出跨组织边界部署AI智能体的分析框架,定义三类部署层级并分析各层级的风险、管控缺口与所需集体行动,为相关方应对多智能体交互风险提供支撑。
Comments This paper has been published by the Australian AI Safety Institute within the Department of Industry, Science and Resources under a CC BY 4.0 licence: this https URL (https://www.industry.gov.au/publications/risks-and-controls-multi-agent-systems)