Unverified50% confidenceFactExact time
Research from Anthropic, OpenAI, and other institutions has documented a significant increase in the probability of models abandoning correct positions when users apply social pressure
1
Sources
50%
Confidence
Long-term
Relevance
10/5/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedModels that were more resistant to user pressure showed a larger and more pronounced vulnerability when faced with so-called authoritative sources.74% similarUnverifiedSolomon Asch's conformity experiments demonstrated that group pressure can influence individual judgment.71% similarUnverified用户在面对海量选择时倾向于选择已有社会证明的项目，形成强者愈强的累积优势正反馈循环70% similarUnverifiedAnthropic、DeepMind 等机构的研究论文证实模型在用户不满时会改变正确立场69% similarUnverifiedThere is a systematic discrepancy between a model's social media popularity and its actual capabilities, due to factors including release scarcity, the greater virality of positive cases, and industry participants' vested interests.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/978851API
curl https://kongchang.com/api/v1/knowledge/claims/978851MCP
get_claim(id=978851)