{"id":9889,"date":"2026-11-05T05:05:51","date_gmt":"2026-11-05T05:05:51","guid":{"rendered":"https:\/\/journals.law.unc.edu\/ncjolt\/?p=9889"},"modified":"2026-11-05T05:09:31","modified_gmt":"2026-11-05T05:09:31","slug":"colorblind-algorithms-after-sffa-race-conscious-compliance-and-the-future-of-civil-rights-in-ai-governance","status":"publish","type":"post","link":"https:\/\/journals.law.unc.edu\/ncjolt\/articles\/volume-28-articles\/colorblind-algorithms-after-sffa-race-conscious-compliance-and-the-future-of-civil-rights-in-ai-governance\/","title":{"rendered":"Colorblind Algorithms After SFFA: Race-Conscious Compliance and the Future of Civil Rights in AI Governance"},"content":{"rendered":"<a href=\"https:\/\/journals.law.unc.edu\/ncjolt\/wp-content\/uploads\/sites\/4\/2026\/11\/COLORBLIND-ALGORITHMS-AFTER-SFFA-Final.pdf\" class=\"pdfemb-viewer\" style=\"\" data-width=\"max\" data-height=\"max\" data-mobile-width=\"500\"  data-scrollbar=\"none\" data-download=\"off\" data-tracking=\"on\" data-newwindow=\"on\" data-pagetextbox=\"off\" data-scrolltotop=\"off\" data-startzoom=\"100\" data-startfpzoom=\"100\" data-toolbar=\"bottom\" data-toolbar-fixed=\"off\">COLORBLIND-ALGORITHMS-AFTER-SFFA-Final<br\/><\/a>\n<p class=\"wp-block-pdfemb-pdf-embedder-viewer\"><\/p>\n\n\n\n<p><em>Artificial intelligence can reproduce racial inequality without ever using race as an explicit input. Hiring models, tenant-screening systems, lending tools, and other automated decision systems can encode racialized patterns through proxies, historical data, and seemingly neutral concepts such as merit, fit, risk, and creditworthiness. Yet the legal response to Students for Fair Admissions v. Harvard (\u201cSFFA\u201d) increasingly treats attention to race itself as suspect. That creates a paradox for civil rights enforcement: institutions may need protected-class data to determine whether an ostensibly race-neutral system discriminates, even as collecting and analyzing those data may be characterized as impermissible, race-conscious decision making. <\/em><\/p>\n\n\n\n<p><em>This Article argues that civil rights law must distinguish race-conscious classification from race-conscious compliance. Classification uses race to determine how an individual will be treated. Compliance uses protected-class information to investigate, validate, audit, and monitor a facially neutral system without allowing race to determine an individual outcome. Existing doctrine already contains the foundations for this distinction, including disparate-impact doctrine, employment-testing law, Ricci v. DeStefano, and Justice Kennedy\u2019s concurrence in Parents Involved. <\/em><\/p>\n\n\n\n<p><em>The Article develops this distinction for algorithmic decision making and extends it to the harder problem of remediation after an audit reveals racial disparities. It proposes a protected, audit-safe harbor for institutions that separates demographic data from operational decision making, examines actual system performance, and uses race-conscious information to identify discrimination without assigning benefits or burdens by race. Without such protection, post-SFFA civil rights law risks creating an incentive for institutions not to measure discrimination at all. In an era when algorithms can reconstruct race without naming it, civil rights law must preserve the ability to see what formally colorblind systems hide.<\/em><\/p>\n\n\n\n<p>PDF: <a href=\"https:\/\/journals.law.unc.edu\/ncjolt\/wp-content\/uploads\/sites\/4\/2026\/11\/COLORBLIND-ALGORITHMS-AFTER-SFFA-Final.pdf\">https:\/\/journals.law.unc.edu\/ncjolt\/wp-content\/uploads\/sites\/4\/2026\/11\/COLORBLIND-ALGORITHMS-AFTER-SFFA-Final.pdf<\/a><\/p>\n\n\n\n<p>Author: Susan Tanner<\/p>\n\n\n\n<p>Volume 28, Issue 1<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Artificial intelligence can reproduce racial inequality without ever using race as an explicit input. Hiring models, tenant-screening systems, lending tools, and other automated decision systems can encode racialized patterns through proxies, historical data, and seemingly neutral concepts such as merit, fit, risk, and creditworthiness. Yet the legal response to Students for Fair Admissions v. Harvard <a href=\"https:\/\/journals.law.unc.edu\/ncjolt\/articles\/volume-28-articles\/colorblind-algorithms-after-sffa-race-conscious-compliance-and-the-future-of-civil-rights-in-ai-governance\/\" class=\"more-link\">&#8230;<\/a><\/p>\n","protected":false},"author":4,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[737,738],"tags":[],"_links":{"self":[{"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/posts\/9889"}],"collection":[{"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/users\/4"}],"replies":[{"embeddable":true,"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/comments?post=9889"}],"version-history":[{"count":1,"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/posts\/9889\/revisions"}],"predecessor-version":[{"id":9891,"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/posts\/9889\/revisions\/9891"}],"wp:attachment":[{"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/media?parent=9889"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/categories?post=9889"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/journals.law.unc.edu\/ncjolt\/wp-json\/wp\/v2\/tags?post=9889"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}