When HR teams say, “The algorithm chose the best applicant,” state lawmakers are increasingly answering with another question: “Who checked the algorithm?” That shift is visible in Illinois, where Public Act 103-0804 took effect on January 1, 2026. The law bars employers from using AI in recruitment, hiring, promotion, training selection, discharge, discipline, tenure, and other employment decisions if the system has the effect of discriminating against protected classes, and it also forbids the use of ZIP codes as a proxy for those classes. The Illinois Department of Human Rights says employers must be transparent about AI use, and the department is still developing implementation rules. (ilga.gov)
Illinois is part of a wider compliance patchwork. In New York City, Local Law 144 has been enforced since July 5, 2023, and employers cannot use an automated employment decision tool unless it has had a bias audit within the past year, audit information is made public, and required notices are given; city guidance also says the notice must come 10 business days before the tool is used. Colorado has gone even broader. Its AI law covers “high-risk” systems that shape consequential decisions, including employment, and although the original law pointed to February 1, 2026, a later measure moved the effective date to June 30, 2026. Colorado’s framework expects risk management, impact assessments, consumer notice, opportunities to correct bad data, and an appeal process with human review when technically feasible. (nyc.gov)
Taken together, these laws suggest that HR cannot outsource judgment to a vendor or hide behind technical jargon. Smart oversight now means mapping every stage where AI influences a decision, checking whether outcomes are unfairly skewed, documenting what data the system uses, banning crude proxies such as ZIP code, and keeping a real human in the loop for important or negative decisions. That approach also fits the federal baseline: the EEOC says existing employment discrimination laws still apply when AI is used, and employers may still have accommodation duties. In other words, the safest hiring algorithm is not the flashiest one. It is the one HR can explain, question, and overrule. (ilga.gov)










