AI Tools for Professions

Free download · print or save as PDF

AI Hiring Compliance Checklist, by State

Every statute, date, and duty below is verified and sourced in our AI hiring compliance guide. Last verified July 2026. Not legal advice — confirm current status with counsel.

The legal map

WhereLawIn forceYour duty as the employer
New York CityLocal Law 144Enforced Jul 5, 2023Independent bias audit (annual) · public summary of results · notice to candidates 10 business days before use. Applies to any AEDT used for NYC-located roles.
IllinoisHB 3773 (Human Rights Act amendment)Effective Jan 1, 2026Civil-rights violation to use AI with a discriminatory effect in employment decisions. Notice required.
IllinoisAI Video Interview ActEffective Jan 1, 2020Notice + explanation of how the AI works + consent BEFORE an AI-analyzed video interview.
ColoradoSB 24-205 (Colorado AI Act)Effective Jun 30, 2026Risk-management duties for high-risk AI systems, incl. hiring. Developer + deployer obligations.
CaliforniaCivil Rights Council ADS regulationsEffective Oct 1, 2025Automated-decision-system rules under FEHA — anti-bias testing and record-keeping expectations.
MarylandHB 12022020Consent required before facial-recognition analysis in interviews.
TexasTRAIGAEffective Jan 1, 2026State AI-governance duties; check current scope for employment uses.
Federal (everywhere)Title VII · ADEA · ADA + EEOC guidance (May 18, 2023)AlwaysThe four-fifths rule applies to an AI screener the same way it applies to a written test. Disparate impact ≠ intent.

The liability anchor

Mobley v. Workday: July 2024 — a federal judge ruled an AI vendor can be sued directly as an agent of the employer; May 2025 — a nationwide age-discrimination collective was certified. Louis v. SafeRent: settled for up to $2.275M (Nov 2024) — the screening-algorithm pattern. The vendor writes "compliant" in the deck and disclaims liability in the contract; the employer holds the bag.

Five questions before you sign with any AI screening vendor

  1. Has this tool had an independent bias audit? Who ran it, when, and on whose data — and can we see the summary?
  2. Does the contract indemnify us, or disclaim vendor liability? (Mobley v. Workday: since July 2024 a vendor can be sued directly as the employer's agent.)
  3. Exactly which decisions does the tool score, rank, or filter — and can we set it to recommend-only?
  4. What data was the model trained on, and is our candidate data used for training?
  5. How do we produce the specific reason behind any candidate's rejection if a regulator or plaintiff asks?

Laws on this page change. We re-verify them and email a short note when they do — no hype, unsubscribe anytime.

lastverified.com · Independent, methodical reviews of AI tools for professional work · Not legal advice