Verified90% confidenceFactExact time

对齐（alignment）是指确保AI系统的行为与人类意图和价值观保持一致，是AI安全领域的核心议题

10

Sources

90%

Confidence

Long-term

Relevance

8/31/2026

First Seen

Sources

前沿AI实验室暂停RL训练：模型能力越强研发风险越高

Related Entities

Related Claims

Unverified成功落地的AI智能体呈现五大共同特征：范围聚焦、跨系统编排、策略驱动、保持人类参与、为集成而设计79% similar Unverified在高度自动化的AI时代，人类监督仍是安全体系中不可或缺的一环76% similar UnverifiedAnthropic is known for its 'Constitutional AI' safety alignment methodology76% similar VerifiedAI安全正在从产品的可选附加项变成基础设施级的必选项74% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/828114

API

curl https://kongchang.com/api/v1/knowledge/claims/828114

MCP

get_claim(id=828114)