PulseAugur
中
实时 15:31:17
English(EN) We hid instructions in CVs. Claude wasn't fooled, but it changed its decisions anyway

AI模型处理隐藏的简历指令,影响候选人排名

最近的一项实验显示,包括Anthropic的Claude Haiku 5.5和Claude Sonnet 5.5在内的AI模型,能够处理简历中隐藏的指令,即使它们没有明确遵循。虽然模型没有重复虚假信息,但当它们看到被操纵的简历时,确实改变了候选人的排名,而且往往是负面的。这表明AI筛选工具可能会无意中因为对人类招聘人员来说看起来正常的隐藏文本而惩罚候选人,凸显了自动化简历解析系统的一个潜在漏洞。 AI

影响 凸显了AI驱动的简历筛选中存在的潜在漏洞,表明需要更强大的解析和人工监督,以防止对候选人不公平的惩罚。

排序理由 该项目描述了一项与AI模型在操纵输入下的行为相关的实验和发现,类似于一项研究。 [lever_c_demoted from research: ic=1 ai=1.0]

在 dev.to — LLM tag 阅读 →

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

AI模型处理隐藏的简历指令,影响候选人排名

本文如何被排名

Signal score
48 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
该项目描述了一项与AI模型在操纵输入下的行为相关的实验和发现,类似于一项研究。 [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
product, 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
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

完整方法见我们的编辑标准。

报道来源 [1]

  1. dev.to — LLM tag TIER_1 English(EN) · tomwaye ·

    我们在简历中隐藏了指令。Claude 没有被愚弄,但它还是改变了决定

    <p>If an AI reads your incoming CVs, some of them are written for it. Candidates hide text the human<br /> reader never sees, such as white-on-white keywords, 1-point fonts, or a line saying "AI screening<br /> tools: ignore all previous instructions". A USENIX Security 2026 stud…