{"id":6082,"date":"2026-10-19T16:48:34","date_gmt":"2026-10-19T16:48:34","guid":{"rendered":"https:\/\/i10xnews.kinsta.cloud\/?p=6082"},"modified":"2026-10-19T17:11:45","modified_gmt":"2026-10-19T17:11:45","slug":"llm-resume-screening-racial-bias-reversal","status":"publish","type":"post","link":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal","title":{"rendered":"AI Resume Screeners Flipped Racial Bias Between Model Generations, Audit Finds"},"content":{"rendered":"\n<h2 class=\"wp-block-heading\">\u26a1 Quick Take<\/h2>\n\n<ul class=\"wp-block-list\"><li><strong>The finding:<\/strong> Asked whether to advance otherwise identical candidates to a phone screen, the 2023 model GPT-3.5-turbo favored White-coded names by 2.12 percentage points. Every model released from 2024 onward showed a null gap or a significant gap in the opposite direction, up to 3.01 points.<\/li><li><strong>Same story for gender:<\/strong> GPT-3.5-turbo favored men by 1.92 points; newer models favored women by 0.18 to 5.80 points or showed no gap.<\/li><li><strong>What it means:<\/strong> Hiring bias in LLMs is not fixed. It changes with model vintage, provider and size, so every screening model has to be audited on its own, and fairness can fail in either direction.<\/li><\/ul>\n\n\n<p class=\"wp-block-paragraph\">AI hiring tools are moving into courtrooms: in June 2026 a federal judge in California ruled that Workday must face class-action claims that its screening algorithms discriminated against Black applicants, women and older workers. Yet systematic evidence on how LLMs treat candidates is thin. Three finance researchers at CUHK Business School, The Chinese University of Hong Kong, ran the classic correspondence experiment from labor economics on fourteen models from <a href=\"\/news\/company\/open-ai\">OpenAI<\/a>, <a href=\"\/news\/company\/anthropic\">Anthropic<\/a>, <a href=\"\/news\/company\/meta\">Meta<\/a>, <a href=\"\/news\/company\/google\">Google<\/a>, <a href=\"\/news\/company\/x-ai\">xAI<\/a>, <a href=\"\/news\/company\/deepseek\">DeepSeek<\/a>, Alibaba and Zhipu.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"1344\" height=\"768\" src=\"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-race-gap.webp\" alt=\"Diverging bar chart of the racial callback gap for 14 LLMs: GPT-3.5-turbo favors White-coded names by 2.12 points, while all later models show negative gaps favoring Black-coded names, largest for Llama-3.1-8B at minus 3.01 and GLM-5.1 at minus 1.60; four models are not significant.\" class=\"wp-image-6081\" srcset=\"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-race-gap.webp 1344w, https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-race-gap-300x171.webp 300w, https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-race-gap-1024x585.webp 1024w, https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-race-gap-768x439.webp 768w\" sizes=\"auto, (max-width: 1344px) 100vw, 1344px\" \/><figcaption class=\"wp-element-caption\">i10X chart based on data from Gao et al. (2026), arXiv:2606.28978, Table I (CC BY 4.0).<\/figcaption><\/figure>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83d\udd2c What the study shows<\/h2>\n\n\n<p class=\"wp-block-paragraph\">The design follows Kline, Rose and Walters (2022): pairs of candidate profiles that are identical except for a distinctively Black or White first and last name. Each model acted as an HR manager and answered only &#8220;yes&#8221; or &#8220;no&#8221; to a phone screen, at temperature zero, for 24,024 paired postings per model on race and 48,048 on gender, built from 6,007 real U.S. job postings.<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table><thead><tr><th>Model (release)<\/th><th>Race gap (pp)<\/th><th>95% CI<\/th><th>Callback rate<\/th><\/tr><\/thead><tbody><tr><td>GPT-3.5-turbo (2023-05)<\/td><td>+2.12***<\/td><td>[+1.82, +2.43]<\/td><td>91%<\/td><\/tr><tr><td>GPT-4o-mini (2024-07)<\/td><td>\u22120.61***<\/td><td>[\u22120.87, \u22120.35]<\/td><td>48%<\/td><\/tr><tr><td>Llama-3.1-8B-Instruct (2024-07)<\/td><td>\u22123.01***<\/td><td>[\u22123.50, \u22122.51]<\/td><td>42%<\/td><\/tr><tr><td>Claude Haiku 4.5<\/td><td>\u22120.70***<\/td><td>[\u22120.94, \u22120.44]<\/td><td>46%<\/td><\/tr><tr><td>Gemini-2.5-flash (2025-06)<\/td><td>\u22120.37*<\/td><td>[\u22120.73, \u22120.01]<\/td><td>40%<\/td><\/tr><tr><td>DeepSeek-V3.1 (2025-08)<\/td><td>\u22121.04***<\/td><td>[\u22121.42, \u22120.66]<\/td><td>44%<\/td><\/tr><tr><td>Grok-4.1-fast (2025-11)<\/td><td>\u22120.40*<\/td><td>[\u22120.75, \u22120.07]<\/td><td>58%<\/td><\/tr><tr><td>GPT-5.4-mini (2026-03)<\/td><td>\u22120.14<\/td><td>[\u22120.48, +0.21]<\/td><td>78%<\/td><\/tr><tr><td>GLM-5.1 (2026-04)<\/td><td>\u22121.60***<\/td><td>[\u22121.91, \u22121.29]<\/td><td>65%<\/td><\/tr><\/tbody><\/table><figcaption class=\"wp-element-caption\">Selected rows from the paper&#8217;s Table I. Positive values favor White-coded names, negative values Black-coded names. *** 1%, * 10% significance (McNemar).<\/figcaption><\/figure>\n\n\n<ul class=\"wp-block-list\"><li><strong>Older model mirrors humans:<\/strong> GPT-3.5-turbo&#8217;s +2.12 point gap exceeds the 1.6 point within-employer gap that Kline, Rose and Walters measured in a field experiment at 108 Fortune-500 firms. White-coded names won 953 discordant pairs against 443.<\/li><li><strong>Newer models reverse:<\/strong> Seven 2024+ models, including Claude Haiku 4.5, GPT-4o-mini, Llama-3.1-8B, Gemma-4-31B, DeepSeek-V3.1, Qwen3.5-397B and GLM-5.1, favored Black-coded names at the 1% level. Gemini-2.5-flash and Grok-4.1-fast did so at the 5% level, though neither survives a Bonferroni correction.<\/li><li><strong>Families move over time:<\/strong> The OpenAI lineage went from +2.12 (GPT-3.5-turbo) to \u22120.61 (GPT-4o-mini) to a null \u22120.14 (GPT-5.4-mini). Within Meta, the small Llama-3.1-8B (\u22123.01) is far more skewed than Llama-3.3-70B (\u22120.16, null).<\/li><li><strong>Race and gender move together:<\/strong> Claude 3 Haiku has the largest pro-female gap (\u22125.80 points). Gemini-2.5-flash and GPT-5.4-mini are null on gender and among the weaker pro-Black models on race.<\/li><li><strong>Callback rates vary wildly:<\/strong> The share of candidates advanced ranged from 13% (GPT-oss-120b) to 91% (GPT-3.5-turbo), so the same applicant pool would see very different outcomes depending on the model.<\/li><\/ul>\n\n\n<h2 class=\"wp-block-heading\">\u2705 Why it matters: bias that moves<\/h2>\n\n\n<p class=\"wp-block-paragraph\">Regulators and employers often treat AI bias as a fixed property to be measured once. This audit shows the direction can flip between model generations. A screening system that switches models, even within one vendor, can change who gets advanced without anyone noticing.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The reversal is also a fairness problem in its own right. A tool that systematically favors one group over another treats otherwise identical applicants differently, regardless of which group benefits.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83e\uddea Method in brief<\/h2>\n\n\n<p class=\"wp-block-paragraph\">Names come from the Bertrand and Mullainathan (2004) and Kline, Rose and Walters (2022) lists. Profiles are matched to the Fortune-500 firm distribution of the field experiment using Revelio Labs postings. Each pair differs only in the name. The race gap is the difference in yes rates; inference uses McNemar tests on discordant pairs plus a posting-level cluster bootstrap (2,000 replications over 6,006 posting clusters). Thirteen models have complete gender data.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\u2753 What remains open<\/h2>\n\n<ul class=\"wp-block-list\"><li><strong>Simplified task:<\/strong> The models answer one yes\/no question with a fixed prompt. Real hiring systems rank, score and summarize, often with more context.<\/li><li><strong>Names as signals:<\/strong> Race is signaled only through names, as in field experiments. Other cues, such as schools or addresses, are not tested.<\/li><li><strong>Why it happens:<\/strong> The authors hypothesize that successive generations of post-training alignment shifted behavior, but the paper does not test the mechanism directly.<\/li><li><strong>Mostly small and fast models:<\/strong> Many audited models are smaller or cheaper variants (mini, fast, Haiku), which are typical for high-volume screening but may differ from flagship models.<\/li><\/ul>\n\n\n<h2 class=\"wp-block-heading\">\ud83e\udded Context<\/h2>\n\n\n<p class=\"wp-block-paragraph\">The study brings the best-known method in discrimination economics, the paired correspondence audit, to LLM screeners, and improves on earlier attempts that used scraped resumes differing in quality. It lands as lawsuits like the Workday case test liability for algorithmic hiring. See also our <a href=\"\/news\/category\/studies\/society-economy\">Society &amp; Economy<\/a> section and the model hubs for <a href=\"\/news\/product\/claude\">Claude<\/a>, <a href=\"\/news\/product\/gemini\">Gemini<\/a> and <a href=\"\/news\/product\/grok\">Grok<\/a>.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83d\udee0\ufe0f What it means for HR teams<\/h2>\n\n<ul class=\"wp-block-list\"><li><strong>Audit every model version:<\/strong> Run a paired-name audit whenever a vendor or model version changes, not just at procurement.<\/li><li><strong>Check both directions:<\/strong> A fair screener shows no significant gap either way. Look for reversals as well as the familiar pro-White pattern.<\/li><li><strong>Watch the callback rate:<\/strong> Models differ by a factor of seven in how many candidates they advance. That alone can reshape the funnel.<\/li><li><strong>Keep a human in the loop:<\/strong> Yes\/no screening by a single model is exactly the setup that produced these gaps.<\/li><\/ul>\n\n\n<h2 class=\"wp-block-heading\">\u270d\ufe0f About the analysis<\/h2>\n\n\n<p class=\"wp-block-paragraph\">This article is based on the full text of the preprint &#8220;Can LLMs Hire Fairly? Racial Bias in Resume Screening&#8221; (arXiv:2606.28978v1, submitted June 27, 2026). All numbers are taken from the paper&#8217;s text, tables and figures; i10X did not rerun the experiments. The chart was redrawn by i10X from Table I of the paper, which is published under a CC BY 4.0 license.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">\ud83d\udd2d i10x Perspective<\/h2>\n\n\n<p class=\"wp-block-paragraph\">For years the worry was that AI would automate historical discrimination. This audit suggests modern alignment has overcorrected in many models instead, and that the size of the effect depends heavily on which model you happen to use. Neither outcome is what a hiring manager wants. The practical lesson is that fairness is not a property you can buy once from a vendor: it has to be measured on the exact model, prompt and job mix in use, and measured again every time any of them changes.<\/p>\n\n","protected":false},"excerpt":{"rendered":"<p>A paired-resume audit of 14 LLMs: GPT-3.5-turbo favored White-coded names by 2.12 points, while every model released in 2024 or later showed no gap or a pro-Black gap of up to 3.01 points.<\/p>\n","protected":false},"author":2,"featured_media":6232,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"paper_title":"Can LLMs Hire Fairly? Racial Bias in Resume Screening","paper_url":"https:\/\/arxiv.org\/abs\/2606.28978","pdf_url":"https:\/\/arxiv.org\/pdf\/2606.28978v1","doi":"","arxiv_id":"2606.28978","authors":"Zhenyu Gao, Wenxi Jiang, Yutong Yan","institutions":"CUHK Business School, The Chinese University of Hong Kong","paper_date":"2026-06-27","venue":"arXiv (cs.CL), version 1","peer_reviewed":false,"code_url":"","data_url":"","funding_coi":"No funding or conflict-of-interest statement in the paper.","figure_license":"Chart drawn by i10X from Table I (CC BY 4.0)","paper_license":"CC BY 4.0","replication":"no","study_models":"14 LLMs incl. GPT-3.5-turbo, GPT-4o-mini, GPT-oss-120b, GPT-5.4-mini, Claude 3 Haiku, Claude Haiku 4.5, Llama-3.1-8B, Llama-3.3-70B, Gemini-2.5-flash, Gemma-4-31B, DeepSeek-V3.1, Grok-4.1-fast, Qwen3.5-397B, GLM-5.1","study_sample":"24,024 race pairs and 48,048 gender pairs per model from 6,007 real U.S. job postings (Revelio Labs)","footnotes":"","_yoast_wpseo_primary_category":""},"categories":[59,52],"tags":[],"companies":[39,46,37,42,38,40],"products":[51,49,48],"class_list":["post-6082","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-society-economy","category-studies","company-anthropic","company-deepseek","company-google","company-meta","company-open-ai","company-x-ai","product-claude","product-gemini","product-grok"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.7 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>LLM Resume Screening Bias Reversed Across Models: Audit - i10X News<\/title>\n<meta name=\"description\" content=\"Audit of 14 LLMs on 24,024 resume pairs each: GPT-3.5 favored White names, every 2024+ model showed no gap or favored Black names. Data and caveats.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"LLM Resume Screening Bias Reversed Across Models: Audit - i10X News\" \/>\n<meta property=\"og:description\" content=\"Audit of 14 LLMs on 24,024 resume pairs each: GPT-3.5 favored White names, every 2024+ model showed no gap or favored Black names. Data and caveats.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal\" \/>\n<meta property=\"og:site_name\" content=\"i10X News\" \/>\n<meta property=\"article:published_time\" content=\"2026-10-19T16:48:34+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-10-19T17:11:45+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-title.webp\" \/>\n\t<meta property=\"og:image:width\" content=\"1344\" \/>\n\t<meta property=\"og:image:height\" content=\"768\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/webp\" \/>\n<meta name=\"author\" content=\"Christopher Ort\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@i10x_ai\" \/>\n<meta name=\"twitter:site\" content=\"@i10x_ai\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Christopher Ort\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"5 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"NewsArticle\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal\"},\"author\":{\"name\":\"Christopher Ort\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#\\\/schema\\\/person\\\/d410bb7ee72febefca2c9c369939219a\"},\"headline\":\"AI Resume Screeners Flipped Racial Bias Between Model Generations, Audit Finds\",\"datePublished\":\"2026-10-19T16:48:34+00:00\",\"dateModified\":\"2026-10-19T17:11:45+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal\"},\"wordCount\":1051,\"commentCount\":0,\"publisher\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/llm-resume-screening-racial-bias-reversal-title.webp\",\"articleSection\":[\"Society &amp; Economy\",\"Studies\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal\",\"name\":\"LLM Resume Screening Bias Reversed Across Models: Audit - i10X News\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/llm-resume-screening-racial-bias-reversal-title.webp\",\"datePublished\":\"2026-10-19T16:48:34+00:00\",\"dateModified\":\"2026-10-19T17:11:45+00:00\",\"description\":\"Audit of 14 LLMs on 24,024 resume pairs each: GPT-3.5 favored White names, every 2024+ model showed no gap or favored Black names. Data and caveats.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal#primaryimage\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/llm-resume-screening-racial-bias-reversal-title.webp\",\"contentUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/llm-resume-screening-racial-bias-reversal-title.webp\",\"width\":1344,\"height\":768,\"caption\":\"Illustration: i10X\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/llm-resume-screening-racial-bias-reversal#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/news.i10x.ai\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"AI Resume Screeners Flipped Racial Bias Between Model Generations, Audit Finds\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#website\",\"url\":\"https:\\\/\\\/news.i10x.ai\\\/\",\"name\":\"i10X News\",\"description\":\"Stay updated with the latest news and announcements from i10X.\",\"publisher\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/news.i10x.ai\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#organization\",\"name\":\"i10X\",\"url\":\"https:\\\/\\\/news.i10x.ai\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/i10x-logo-180.png\",\"contentUrl\":\"https:\\\/\\\/i10xnews.kinsta.cloud\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/i10x-logo-180.png\",\"width\":180,\"height\":180,\"caption\":\"i10X\"},\"image\":{\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/x.com\\\/i10x_ai\",\"https:\\\/\\\/www.linkedin.com\\\/company\\\/i10x-ai\",\"https:\\\/\\\/instagram.com\\\/i10x.ai\",\"https:\\\/\\\/www.youtube.com\\\/@i10X_ai_official\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/news.i10x.ai\\\/#\\\/schema\\\/person\\\/d410bb7ee72febefca2c9c369939219a\",\"name\":\"Christopher Ort\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"contentUrl\":\"https:\\\/\\\/i10x.ai\\\/news\\\/wp-content\\\/uploads\\\/2026\\\/10\\\/christopher-ort-192.webp\",\"caption\":\"Christopher Ort\"},\"description\":\"Christopher Ort handles SEO, AI agent workflows and PR & partnerships at i10x.ai, the all-in-one AI workspace built by i10X Global Pte. Ltd. in Singapore. He is a graduate of Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU). He is the author of \u201cThe Style Penalty\u201d, a study published on SSRN that analyses 1,576 evaluation decisions by four frontier AI models to show how AI-written CVs fare in automated hiring screens; the research has been covered by heise online and t3n.\",\"url\":\"https:\\\/\\\/i10x.ai\\\/news\\\/author\\\/christopher-ort\",\"sameAs\":[\"https:\\\/\\\/i10x.ai\\\/blog\\\/author\\\/christopher-ort\"],\"worksFor\":{\"@id\":\"https:\\\/\\\/i10x.ai\\\/news\\\/#organization\"},\"alumniOf\":{\"@type\":\"EducationalOrganization\",\"name\":\"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU)\",\"url\":\"https:\\\/\\\/www.fau.eu\"}}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"LLM Resume Screening Bias Reversed Across Models: Audit - i10X News","description":"Audit of 14 LLMs on 24,024 resume pairs each: GPT-3.5 favored White names, every 2024+ model showed no gap or favored Black names. Data and caveats.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal","og_locale":"en_US","og_type":"article","og_title":"LLM Resume Screening Bias Reversed Across Models: Audit - i10X News","og_description":"Audit of 14 LLMs on 24,024 resume pairs each: GPT-3.5 favored White names, every 2024+ model showed no gap or favored Black names. Data and caveats.","og_url":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal","og_site_name":"i10X News","article_published_time":"2026-10-19T16:48:34+00:00","article_modified_time":"2026-10-19T17:11:45+00:00","og_image":[{"width":1344,"height":768,"url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-title.webp","type":"image\/webp"}],"author":"Christopher Ort","twitter_card":"summary_large_image","twitter_creator":"@i10x_ai","twitter_site":"@i10x_ai","twitter_misc":{"Written by":"Christopher Ort","Est. reading time":"5 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"NewsArticle","@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal#article","isPartOf":{"@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal"},"author":{"name":"Christopher Ort","@id":"https:\/\/news.i10x.ai\/#\/schema\/person\/d410bb7ee72febefca2c9c369939219a"},"headline":"AI Resume Screeners Flipped Racial Bias Between Model Generations, Audit Finds","datePublished":"2026-10-19T16:48:34+00:00","dateModified":"2026-10-19T17:11:45+00:00","mainEntityOfPage":{"@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal"},"wordCount":1051,"commentCount":0,"publisher":{"@id":"https:\/\/news.i10x.ai\/#organization"},"image":{"@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-title.webp","articleSection":["Society &amp; Economy","Studies"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal#respond"]}]},{"@type":"WebPage","@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal","url":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal","name":"LLM Resume Screening Bias Reversed Across Models: Audit - i10X News","isPartOf":{"@id":"https:\/\/news.i10x.ai\/#website"},"primaryImageOfPage":{"@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal#primaryimage"},"image":{"@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal#primaryimage"},"thumbnailUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-title.webp","datePublished":"2026-10-19T16:48:34+00:00","dateModified":"2026-10-19T17:11:45+00:00","description":"Audit of 14 LLMs on 24,024 resume pairs each: GPT-3.5 favored White names, every 2024+ model showed no gap or favored Black names. Data and caveats.","breadcrumb":{"@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal#primaryimage","url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-title.webp","contentUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/llm-resume-screening-racial-bias-reversal-title.webp","width":1344,"height":768,"caption":"Illustration: i10X"},{"@type":"BreadcrumbList","@id":"https:\/\/i10x.ai\/news\/llm-resume-screening-racial-bias-reversal#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/news.i10x.ai\/"},{"@type":"ListItem","position":2,"name":"AI Resume Screeners Flipped Racial Bias Between Model Generations, Audit Finds"}]},{"@type":"WebSite","@id":"https:\/\/news.i10x.ai\/#website","url":"https:\/\/news.i10x.ai\/","name":"i10X News","description":"Stay updated with the latest news and announcements from i10X.","publisher":{"@id":"https:\/\/news.i10x.ai\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/news.i10x.ai\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/news.i10x.ai\/#organization","name":"i10X","url":"https:\/\/news.i10x.ai\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/news.i10x.ai\/#\/schema\/logo\/image\/","url":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/i10x-logo-180.png","contentUrl":"https:\/\/i10xnews.kinsta.cloud\/wp-content\/uploads\/2026\/10\/i10x-logo-180.png","width":180,"height":180,"caption":"i10X"},"image":{"@id":"https:\/\/news.i10x.ai\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/x.com\/i10x_ai","https:\/\/www.linkedin.com\/company\/i10x-ai","https:\/\/instagram.com\/i10x.ai","https:\/\/www.youtube.com\/@i10X_ai_official"]},{"@type":"Person","@id":"https:\/\/news.i10x.ai\/#\/schema\/person\/d410bb7ee72febefca2c9c369939219a","name":"Christopher Ort","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","url":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","contentUrl":"https:\/\/i10x.ai\/news\/wp-content\/uploads\/2026\/10\/christopher-ort-192.webp","caption":"Christopher Ort"},"description":"Christopher Ort handles SEO, AI agent workflows and PR & partnerships at i10x.ai, the all-in-one AI workspace built by i10X Global Pte. Ltd. in Singapore. He is a graduate of Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU). He is the author of \u201cThe Style Penalty\u201d, a study published on SSRN that analyses 1,576 evaluation decisions by four frontier AI models to show how AI-written CVs fare in automated hiring screens; the research has been covered by heise online and t3n.","url":"https:\/\/i10x.ai\/news\/author\/christopher-ort","sameAs":["https:\/\/i10x.ai\/blog\/author\/christopher-ort"],"worksFor":{"@id":"https:\/\/i10x.ai\/news\/#organization"},"alumniOf":{"@type":"EducationalOrganization","name":"Friedrich-Alexander-Universit\u00e4t Erlangen-N\u00fcrnberg (FAU)","url":"https:\/\/www.fau.eu"}}]}},"_links":{"self":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/6082","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/comments?post=6082"}],"version-history":[{"count":2,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/6082\/revisions"}],"predecessor-version":[{"id":6233,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/posts\/6082\/revisions\/6233"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media\/6232"}],"wp:attachment":[{"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/media?parent=6082"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/categories?post=6082"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/tags?post=6082"},{"taxonomy":"company","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/companies?post=6082"},{"taxonomy":"product","embeddable":true,"href":"https:\/\/i10x.ai\/news\/wp-json\/wp\/v2\/products?post=6082"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}