[KongchangAI]
Unverified50% confidenceOpinionExact time

多智能体系统存在委托链风险、责任分散、激励不对齐三类结构性对齐风险,推动AI安全界强调系统级对齐而非单模型对齐

1
Sources
50%
Confidence
Long-term
Relevance
7/8/2026
First Seen

Sources

Related Claims

Cite This Claim

Stable URI
https://kongchang.com/claim/230993
API
curl https://kongchang.com/api/v1/knowledge/claims/230993
MCP
get_claim(id=230993)