Where to use AI in hiring — and where not — CHRO & People Compass

When it comes up

When recruiters ask for AI tools, when application volume spikes, when an HR-tech vendor adds AI screening to your applicant tracking system, or when you hire in places with AI hiring laws.

Your options

AI for admin only

Fits when: You want quick, low-risk wins: scheduling, candidate Q&A, drafting job ads and interview guides.

Watch out: Recruiters may still use unapproved screening tools on their own; provide approved helpers and a clear policy.

AI-assisted screening with human decisions

Fits when: High volumes, clear job-related criteria, a vendor that supports bias audits, and recruiters trained to review every output.

Watch out: Bias, notice and audit duties (NYC LL144, Illinois, EU high-risk from December 2027); automation bias — reviewers rubber-stamping scores.

Automated rejection or ranking without review

Fits when: Rarely appropriate; only for objective, legally required knock-out criteria (for example a required licence), applied transparently.

Watch out: Highest legal and reputational risk; hard to defend if outcomes differ by group; often prohibited or restricted.

How to decide

List every hiring step and the AI already in it (including features switched on in your ATS).
Sort each use into predict, draft or act. Admin helpers (draft, act on logistics) are low risk; anything that scores, ranks or filters people is high-control.
Map the law by location: NYC Local Law 144 (bias audit within a year, published summary, 10 business days' notice), Illinois (video-interview consent; HB 3773 notice and anti-discrimination from 2026), EU AI Act high-risk from 2 December 2027, plus general anti-discrimination law everywhere.
Set the rule: AI can sort the admin; people make every decision about a candidate. Allow AI-assisted screening only with bias testing, notice, a human review and an alternative process.
Pilot with a baseline (time to hire, selection rates by group, candidate experience) and review at 90 days.

Trade-offs

Admin-only AI is safe and fast but leaves screening workload untouched; AI-assisted screening saves the most recruiter time but brings legal duties and bias risk; automated decisions save the most time and carry the most risk.

Real examples

The U.S. EEOC said in 2023 that iTutorGroup's application software automatically rejected female applicants aged 55 or older and male applicants aged 60 or older; the company agreed to pay $365,000. Source: U.S. EEOC — iTutorGroup to pay $365,000 to settle EEOC discriminatory hiring suit (Sep 11, 2023) ↗
In May 2025 a U.S. federal court preliminarily certified a nationwide age-discrimination collective action in Mobley v. Workday over AI-based applicant recommendations, showing that vendors as well as employers can face claims. Source: Davis Wright Tremaine — Federal court allows age-discrimination collective action over AI hiring tools (Mobley v. Workday, May 2025) ↗
Chipotle's conversational hiring assistant (2024) schedules interviews, answers questions and sends offers — but only to candidates selected by managers: admin is automated, the decision stays human. Source: Chipotle — Chipotle introduces new AI hiring platform to support its accelerated growth (Oct 22, 2024) ↗

What to check first

Every AI feature in your ATS and assessments, and which steps it affects.
The locations you hire in and the AI hiring laws that apply there.
Whether your vendors can provide bias audit results and support your own audit.
Your current selection rates by group, as a baseline.

Learn and use

Sources

Last reviewed 2026-10-01.