Unverified50% confidenceFactExact time
If training corpora contain gender, racial, or cultural bias, models will inherit these biases.
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
UnverifiedThe gender bias in word embedding models originates from training corpora (such as the Google News corpus) reflecting pre-existing human social attitudes, rather than being artificially introduced.76% similarUnverified来自不同厂商、经历不同训练路径的模型持有不同的先验偏见，从而在交叉审查中形成真正的视角差异73% similarUnverifiedexposure bias的根源在于训练与推理的分布不匹配，训练时使用teacher forcing策略而推理时模型依赖自身生成的token72% similarUnverified训练数据构成决定模型世界观边界的现象在数字人文领域被称为'语料库偏差'（Corpus Bias）71% similarUnverified商用人脸识别系统在不同种族和性别群体间存在显著准确率差异，深肤色人群和女性的误识率明显偏高，根源在于训练数据构成不均衡70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/751614API
curl https://kongchang.com/api/v1/knowledge/claims/751614MCP
get_claim(id=751614)