Expired50% confidenceOpinionExact time
An Anthropic co-founder acknowledged that what the emotion-like signals discovered inside AI models mean remains unclear and requires ongoing discernment.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/2/2026
First Seen
Valid until: 9/30/2026(expired)
Sources
Related Claims
UnverifiedAnthropic researchers have discovered emotion-like signals inside AI models85% similarUnverifiedThe Anthropic research team discovered internal states within AI models that are functionally analogous to happiness, satisfaction, fear, sadness, and unease.75% similarUnverifiedAI直接描述情绪在统计上是更"安全"、更高频的表达选择，符合模型对"好写作"的概率预测74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/60522API
curl https://kongchang.com/api/v1/knowledge/claims/60522MCP
get_claim(id=60522)