Unverified50% confidenceFactExact time

Anthropic在其「Constitutional AI」框架中强调AI系统应诚实、无害、对用户有益

1

Sources

50%

Confidence

Long-term

Relevance

7/12/2026

First Seen

Sources

Claude新功能Reflect：可视化数据背后的AI依赖养成逻辑

rss7/9/2026

Related Claims

UnverifiedConstitutional AI has achieved significant results in improving the model's helpfulness, harmlessness, and honesty (the HHH principles)81% similar VerifiedAnthropic提出了宪法AI（Constitutional AI）和可解释性研究等安全框架80% similar

Cite This Claim

Stable URI

https://kongchang.com/claim/493212

API

curl https://kongchang.com/api/v1/knowledge/claims/493212

MCP

get_claim(id=493212)