Unverified50% confidenceFactExact time
Content moderation models assign higher toxicity scores to text containing features of African American Vernacular English (AAVE).
1
Sources
50%
Confidence
Long-term
Relevance
8/15/2026
First Seen
Sources
AI内容审核的偏见与双标困境：技术根源与解决路径
Related Claims
Unverified大语言模型在生成摘要时天然倾向于反映评论中的多数意见，可能将低频的安全警告边缘化61% similarUnverifiedIn production environments, users may input content containing ambiguity, typos, mixed multilingual text, or even malicious injections (Prompt Injection).56% similarVerifiedTypical characteristics of text output by mainstream large language models include well-structured sentence patterns, clear paragraph organization, frequent use of emphatic vocabulary, and explicit logical chains — a phenomenon researchers refer to as "stylistic bias."56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/750742API
curl https://kongchang.com/api/v1/knowledge/claims/750742MCP
get_claim(id=750742)