Introduction
Five years ago, AI hiring regulation was a New York City experiment. In 2026 it is a patchwork — five US jurisdictions with enforceable rules, the EU AI Act's employment provisions arriving in August, and a stack of pending bills that could double the map within two years. Employers using AI screening or interviewing now carry compliance obligations most of them have never read.
This page maps every US state with an enacted AI hiring law, explains what each one actually requires, and flags what is coming next. It is written for employers, but candidates will find their rights here too. For the federal picture, see our EEOC and Local Law 144 compliance guide, and for Europe, our EU AI Act briefing.
Quick Answer: Yes, AI interviewing is legal everywhere in the US in 2026, but five jurisdictions regulate it. New York City requires an annual bias audit and candidate notice. Illinois requires consent and deletion rights for AI video interviews and now covers all AI employment decisions. Colorado, California, and Maryland add disclosure, anti-discrimination, and facial-analysis consent duties.
The compliance map at a glance
| Jurisdiction | Law | Core requirement | Penalty exposure |
|---|---|---|---|
| New York City | Local Law 144 | Annual independent bias audit, public summary, 10-business-day candidate notice | $500 to $1,500 per violation, per day |
| Illinois | AI Video Interview Act + 2026 IHRA amendment | Consent before AI video analysis, explanation of how it works, deletion within 30 days on request, and as of January 2026, no discriminatory AI use across all employment decisions | Civil rights enforcement and private complaints |
| Colorado | SB 24-205, the Colorado AI Act | Duty of reasonable care for high-risk AI including hiring, impact assessments, candidate disclosure, adverse-action explanations | Attorney General enforcement, up to $20,000 per violation |
| California | FEHA automated-decision regulations | Anti-discrimination testing duties for automated decision systems, recordkeeping, vendor due diligence | FEHA discrimination liability |
| Maryland | HB 1202 | Written consent before facial recognition analysis of video interviews | Employment law liability |
Three patterns cut across all five. Disclosure to candidates is becoming universal. Bias testing obligations sit with the employer, not the vendor. And "our vendor handles compliance" is not a defense any regulator has accepted.
New York City: Local Law 144
The law that started the US wave has been enforceable since July 2023. If you use an automated employment decision tool to screen candidates for jobs tied to New York City, you must commission an independent bias audit within one year of each use, publish a summary on your website, and notify candidates at least 10 business days before the tool evaluates them. Candidates must be told what data the tool uses and can request an alternative process.
The fines look modest at $500 to $1,500 per violation, but each day of non-compliant use and each affected candidate can count separately, which is how a quiet gap becomes a six-figure problem. The audit must be genuinely independent — a vendor auditing its own tool does not qualify.
Illinois: from video interviews to all AI decisions
Illinois has regulated AI-analyzed video interviews since 2020 under the Artificial Intelligence Video Interview Act. Before an employer subjects a video interview to AI analysis, the candidate must be notified, told how the analysis works, and give consent. Candidates can demand deletion of the video within 30 days, and sharing is restricted to people whose expertise is needed to evaluate the candidate.
The bigger shift arrived January 1, 2026, when an amendment to the Illinois Human Rights Act extended anti-discrimination rules to AI use across all employment decisions — hiring, promotion, discipline, and termination. Employers must ensure their AI tools do not produce discriminatory outcomes and cannot use zip code as a proxy for protected characteristics.
Colorado: the broadest state law
The Colorado AI Act, with obligations taking effect in 2026 after a delayed start, is the first comprehensive US state AI law. Hiring tools fall squarely in its high-risk category. Employers deploying high-risk AI owe a duty of reasonable care to avoid algorithmic discrimination, must run impact assessments, must tell candidates when AI materially influences a decision, and must explain adverse decisions with a path to human review.
The practical ask for TA teams is documentation. If you cannot produce an impact assessment and vendor due-diligence file for each AI hiring tool, you are carrying exposure you cannot see.
California: regulation through existing law
California chose to regulate AI hiring through its existing Fair Employment and Housing Act rather than a new AI statute. Regulations effective October 2025 make clear that automated decision systems used in hiring must meet anti-discrimination standards, that employers must keep records of how these systems are used and tested, and that vendor selection itself is a due-diligence act. Expect this framework to be the template other states copy.
Maryland: the facial analysis consent rule
Since 2020, Maryland has required written consent before an employer uses facial recognition technology in a video interview. It is a narrower law than Illinois', but it matters for any platform that touches visual analysis — and it is why credible vendors removed facial analysis from their models years ago.
What employers should actually do
- Inventory every AI tool in your funnel. Screening, scheduling, assessment, sourcing. You cannot comply with what you have not mapped.
- Match tools to jurisdictions. Hiring remotely across states means the strictest applicable law usually becomes your default standard.
- Demand bias-audit documentation before signature. A vendor with auditable, rubric-based scoring can produce this. A vendor whose scoring is a black box cannot. This is one reason our evaluation framework weights scoring explainability so heavily, and why auditable scoring is a core differentiator in our Tenzo AI review — rubric-level evidence is exactly what these laws ask for.
- Write the candidate notice now. Disclosure is required in NYC, Illinois, Colorado, and Maryland in different forms. One clear notice covers all of them. Our guide to communicating AI screening to candidates has the templates.
- Audit annually, not once. NYC requires it yearly. Treat that as the floor everywhere. The bias audit guide walks through the process.
For a structured readiness review, the compliance and audit-readiness use case page collects the relevant tools and research.
What candidates should know
If an employer uses AI to evaluate you in a regulated jurisdiction, you generally have the right to be told before it happens, to know what is being analyzed, and in some places to request deletion or a human alternative. If you were never told AI would evaluate your interview and you applied in New York City, Illinois, Colorado, or Maryland, the employer may owe you more than an apology. Our guide to how AI interviews actually work explains what happens on the other side of the camera.
FAQ
Is it legal for AI to reject a job candidate?
Yes, it is legal in every US state for AI to screen or reject candidates, provided the employer complies with anti-discrimination law and any applicable AI-specific rules. What is illegal everywhere is a tool that produces discriminatory outcomes, and in New York City, Illinois, Colorado, California, and Maryland there are additional audit, consent, or disclosure duties.
Do employers have to tell candidates that AI is being used?
In New York City, Illinois, Colorado, and Maryland, yes, in different forms. NYC requires notice 10 business days before an automated evaluation. Illinois requires notice and consent before AI video analysis. Elsewhere disclosure is not yet legally required, but candidate experience data strongly favors telling people, and more states are adding the requirement each session.
What is a bias audit under NYC Local Law 144?
A bias audit is an independent statistical test of an automated employment decision tool that measures selection rates and impact ratios across race, sex, and intersectional categories. The audit must be conducted by an independent auditor within one year of the tool's use, and a summary of results must be published on the employer's website.
Who is liable when an AI hiring tool discriminates, the vendor or the employer?
The employer. Every enacted US law and the EEOC's 2023 guidance put the duty on the employer using the tool. Vendor contracts may include indemnification, but regulators and courts hold the deploying employer responsible, which is why vendor due diligence and auditable scoring matter.
Does the EU AI Act apply to US companies?
It applies to any company whose AI hiring system is used to evaluate people in the EU, regardless of where the company is headquartered. Employment and worker management are classified as high-risk uses, with obligations phasing in through August 2026. US companies hiring in Europe should read our EU AI Act briefing before their next requisition opens there.
Which states are likely to pass AI hiring laws next?
Bills are active in New Jersey, Massachusetts, Connecticut, and several other states, most modeled on either the Colorado comprehensive approach or the California regulation-through-existing-law approach. We update this map quarterly as sessions close.
Evaluating AI recruiting software?
Download the vendor scorecard template and RFP question bank — structured tools for every stage of the buying process.
Vendor ScorecardAbout the author
Editorial Research Team
Platform Evaluation and Buyer Guides
Practitioners with direct experience in enterprise TA leadership, HR technology procurement, and staffing operations. All buyer guides apply our published 100-point evaluation rubric.
Free Consultation
Get a shortlist built for your ATS and volume
Our research team builds custom shortlists based on your ATS, hiring volume, and specific requirements. No cost, no vendor access to your contact information.
Related Articles
AI Hiring Compliance 2026: EEOC Guidance, NYC Local Law 144, and the Illinois AIVIA
AI hiring compliance guide for 2026. EEOC guidance, NYC Local Law 144, and Illinois AIVIA requirements — what enterprise TA teams must document and audit.
AI Recruiting Statistics: 40 Numbers That Matter in 2026
Forty verified AI recruiting statistics for 2026 covering adoption, screening, candidate experience, cost, and compliance, each with a source.
Amazon Interview Process 2026: What to Expect at Every Step
The Amazon interview process in 2026 explained step by step — online assessments, phone screens, the Loop, Bar Raisers, Leadership Principles, and timelines.
How to Pass an AI Interview: 2026 Prep Guide
How to pass an AI interview in 2026 — what the software actually measures, prep by format, STAR answers that score, tech setup, and the mistakes to avoid.
AI Interview Questions and Answers for 2026
The AI interview questions employers actually ask in 2026, organized by role, with STAR answer frameworks and the follow-up questions AI interviewers probe.
Are AI Interviews Recorded? 15 Questions Candidates Ask
Are AI interviews recorded, do they watch your face, can they detect lies? Honest answers to the 15 questions candidates actually ask about AI interviews.
_1769007509876-Dl4rMdXg.avif)