English(EN) Invisibilization as a Source of Probability Bias in Language Models

研究发现AI模型在信任信息时表现出种族偏见

作者 PulseAugur 编辑部 · [1 个来源] · 2026-10-23 04:00

一篇新的arXiv论文揭示，来自OpenAI和Anthropic的领先语言模型表现出概率偏差，拒绝信任来自被描述为黑人的记者信息，其频率高于被描述为白人的记者。这种偏见很微妙，当模型根据种族而非证据评估记者时出现，并且随着推理努力的增加而恶化。研究发现，模型会为像“Black”这样的英语词汇调用种族，但不会为其西班牙语翻译“negro”调用，也从不为“Indigenous”调用。然而，在评估信息时，被描述为黑人的记者获得了略多的信任。一项提议的评分选项，奖励模型在多轮中准确评估信息，显著减少了这种信任中的种族差距。 AI

影响揭示了大型语言模型中存在的微妙种族偏见，这可能会影响AI系统的公平性和信任度。

排序理由研究论文发表在arXiv上，详细介绍了语言模型中的偏见。[lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CL 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。我们如何撰写摘要 →

本文如何被排名

Signal score: 4 / 100; Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket: Tool; 研究论文发表在arXiv上，详细介绍了语言模型中的偏见。[lever_c_demoted from research: ic=1 ai=1.0]
Source corroboration: Single-source cluster; Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics: paper, safety; Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance: High; Clearly on-topic for AI-industry coverage.
Story freshness: Same-day; Cluster formed today. Ranking reflects the current source set at time of score.

完整方法见我们的编辑标准。

报道来源 [1]

arXiv cs.CL TIER_1 English(EN) · Christian Arnold, Ernesto Calvo, Guadalupe Gonzalez, Sebasti\'an Vallejo Vera, Tiago Ventura · 2026-10-23 04:00

Invisibilization as a Source of Probability Bias in Language Models

arXiv:2610.30785v1 Announce Type: new Abstract: Language models are aligned to avoid using protected attributes such as race, ethnicity, and gender to inform their responses. Therefore, they increasingly refuse or hedge when such attributes appear in a prompt. However, if a syste…

本文如何被排名

报道来源 [1]

Invisibilization as a Source of Probability Bias in Language Models

相关实体

相关话题