Using AI without creating unfair outcomes
AI can match candidates to roles, parse resumes into structured profiles, summarize interviews, answer employee questions about policy and forecast attrition. These uses are valuable but sensitive: models trained on past hiring can copy past bias, and a candidate rejected by a black box has little recourse and may have legal rights. Generated explanations and structured scorecards also help hiring managers compare candidates on job-related criteria instead of gut feeling, which supports consistency across interviewers.
We build with safeguards: feature choices reviewed for proxies of protected traits, bias testing across groups, explanations a recruiter can read, and human decisions on every rejection. Regulations such as NYC Local Law 144 and the EU AI Act treat hiring tools as high-risk, so audit evidence is part of the product. Candidates should also be told when automation is used and be offered a route to human review, which is increasingly a legal requirement and always good practice.
