50-State AI Employment Legislation Survey
The most thorough multi-state survey in the practitioner space. Tracks enacted laws, pending bills, and regulatory developments across all 50 states with employer-side analysis. Cited by most other firms.
The Brief On
U.S. AI Employment Laws
On laws regulating use of AI tools in the workplace.
Updated April 26, 2026
TL;DR
State and city laws are filling a federal vacuum.
No federal statute or preemption.
Congress has not passed an AI employment law. President Trump's December 11, 2025 executive order ("Ensuring a National Policy Framework for AI") directs the DOJ to establish an "AI Litigation Task Force" charged with challenging state AI laws, and specifically names the Colorado AI Act as a problematic example. The EO itself does not preempt any state law. On April 24, 2026, the DOJ filed a statement of interest joining xAI's pending challenge to the Colorado AI Act, the first publicly disclosed Task Force action. Practitioners advise: comply with state laws, don't wait.
Five enacted laws, three already live.
NYC LL 144 (2023), California ADS regs (Oct 2025), Illinois HB 3773 (Jan 2026), and Texas TRAIGA (Jan 2026) are in effect. Colorado AI Act was delayed from Feb 2026 to June 30, 2026 by SB 25B-004.
Two shared requirements across enacted laws.
Every law touches some form of (1) notice / transparency to employees and applicants, and (2) anti-discrimination obligation, but the standard varies sharply: most laws are effects-based / disparate impact; Texas is intent-based; NYC is process-based (mandatory bias audit).
Multi-state trackers
State-by-State AI Legislation Snapshot
A live state tracker with side-by-side comparison of enacted and pending AI legislation. Strong on regulatory scope and definitional differences across states; useful as a sanity check against Littler.
Recent articles
Enacted laws
State AI Laws — Where Are They Now?
Comprehensive April 2026 status check on every enacted state AI regime: effective dates, current enforcement posture, and how the federal EO is (and isn't) reshaping state-level compliance.
Several State AI Laws Set to Go into Effect in 2026
Walks through the four state laws plus NYC LL 144 that are now operative, with the federal EO as backdrop. Useful framing for non-lawyer stakeholders.
New State AI Laws Are Effective on January 1, 2026, But a New Executive Order Signals Disruption
Big-firm read on how the Dec 11 EO interacts with the Jan 1 effective dates for IL HB 3773 and TX TRAIGA. Practical compliance posture during federal-state tension.
Pending bills
Proposed State AI Law Update — April 20, 2026
Biweekly digest of every active state AI bill with movement that cycle. Best source for "what moved this week" across pending state AI legislation.
State AI Legislation Tracker 2026: All 50 States
Bill-level tracker focused on legislation likely to advance. Tags by category and shows committee status. Strongest for spotting which pending bills are positioned to actually pass.
Enacted laws
| Jurisdiction | Law | Scope | Carve-outs |
|---|---|---|---|
| Illinois | HB 3773 Illinois Department of Human Rights guidance·Public Act 103-0804·Draft IDHR Subpart J Rules | Any machine-based system that infers outputs influencing physical or virtual environments. Generative AI explicitly included. PA 103-0804 § 1-103 | Word processors, spreadsheets, spell-checkers, calculators, databases, firewalls, anti-virus, and AI features not used to "make, influence, or facilitate" a covered decision. Draft Subpart J § 2520.110 |
| New York City | Local Law 144 Department of Consumer and Worker Protection guidance·6 RCNY § 5-300·NYC Admin Code Subchapter 25 | No standalone definition of "artificial intelligence." LL 144 defines "automated employment decision tool" (AEDT): any computational process derived from machine learning, statistical modeling, data analytics, or artificial intelligence that issues a simplified output (score, classification, or recommendation) used to substantially assist or replace discretionary decision-making in employment. NYC Admin Code § 20-870 | Tools that don't materially impact natural persons or substantially assist/replace discretionary decision-making: junk email filters, firewalls, antivirus software, calculators, spreadsheets, databases, and other compilations of data. NYC Admin Code § 20-870 |
| Colorado | AI Act (SB 24-205) Colorado Attorney General guidance·SB 24-205 | "High-risk," meaning systems that make or substantially factor in consequential decisions about consumers (including employment). SB 24-205 § 6-1-1701 | Lower-risk uses outside the consequential-decision list. SB 24-205 § 6-1-1703 |
| California | Automated-Decision Systems (ADS) Regulations under the Fair Employment and Housing Act (FEHA) California Civil Rights Department announcement·Final FEHA ADS regulations | A machine-based system that infers, from the input it receives, how to generate outputs. Outputs can include predictions, content, recommendations, or decisions. 2 CCR § 11008(c) | Limited statutory carve-outs; reg-level guidance. 2 CCR § 11008(b) |
| Texas | TRAIGA (HB 149) HB 149 bill page·HB 149 text | Machine-based system that infers outputs from inputs to influence physical or virtual environments. Generative AI included. HB 149 § 552.001 | Statutory scope is broad (covers any developer or deployer doing business in Texas, producing a product or service Texans use, or operating AI in Texas), but employer-specific obligations are limited: no notice, audit, or impact-assessment mandates for private employers. HB 149 §§ 552.002, 552.052 |
| Jurisdiction | Law | Timing | Posting location | Required content |
|---|---|---|---|---|
| Illinois | HB 3773 Illinois Department of Human Rights guidance·Public Act 103-0804·Draft IDHR Subpart J Rules | Annual to current employees; within 30 days of new or substantially-updated AI; in every job posting. Draft Subpart J § 2520.910(b) | Four channels (cumulative): handbook, premises poster, intranet or external site (with conspicuous homepage link), job postings. Draft Subpart J § 2520.910(c)–(d) | Seven elements: product name and vendor; covered decision; data categories; positions affected; HR contact; accommodation rights; statutory anti-discrimination language. Draft Subpart J § 2520.910(e) |
| New York City | Local Law 144 Department of Consumer and Worker Protection guidance·6 RCNY § 5-300·NYC Admin Code Subchapter 25 | At least 10 business days before use of an automated employment decision tool (AEDT). NYC Admin Code § 20-871 | Direct notice to the candidate, plus public posting of the annual bias-audit summary on the employer's website. 6 RCNY §§ 5-301, 5-303 | Disclosure of automated employment decision tool use; data categories collected; bias-audit results. 6 RCNY § 5-303 |
| Colorado | AI Act (SB 24-205) Colorado Attorney General guidance·SB 24-205 | Pre-decision (before the consequential decision is made). SB 24-205 § 6-1-1703 | Notice to the consumer plus risk-management-program disclosures from the deployer. SB 24-205 § 6-1-1703 | System purpose; nature of the consequential decision; risk-management information. SB 24-205 § 6-1-1703 |
| California | Automated-Decision Systems (ADS) Regulations under the Fair Employment and Housing Act (FEHA) California Civil Rights Department announcement·Final FEHA ADS regulations | Pre-use in some contexts; otherwise context-dependent. FEHA ADS regulations | No enumerated channel mandate. FEHA ADS regulations | No enumerated content elements. FEHA ADS regulations |
| Texas | TRAIGA (HB 149) HB 149 bill page·HB 149 text | Not applicable to employers. HB 149 | Not applicable to employers. HB 149 | No notice, disclosure, or transparency obligation on private or public employers under TRAIGA. The statute's notice rule covers government agencies dealing with non-employment "consumers"; "consumer" is statutorily defined to exclude employment. HB 149 §§ 552.001(7), 552.054 |
| Jurisdiction | Law | Standard | Liability trigger | Defense / safe harbor | Notable |
|---|---|---|---|---|---|
| Illinois | HB 3773 Illinois Department of Human Rights guidance·Public Act 103-0804·Draft IDHR Subpart J Rules | Effects-based. PA 103-0804 § 2-102(L) | AI use that has the effect of subjecting employees to discrimination on any class protected by the Illinois Human Rights Act (IHRA). Intent not required. PA 103-0804 § 2-102(L)·775 ILCS 5/1-103(Q) | None statutory. Privileged bias audits as a practitioner-recommended de facto defense. PA 103-0804 (no statutory safe harbor) | Bans ZIP code as a proxy for protected classes. 19+ protected classes. PA 103-0804 § 2-102(L)·775 ILCS 5/1-103(Q) |
| New York City | Local Law 144 Department of Consumer and Worker Protection guidance·6 RCNY § 5-300·NYC Admin Code Subchapter 25 | Process-based. NYC Admin Code § 20-871 | Failure to conduct or post the annual bias audit, or to provide candidate notice. 6 RCNY §§ 5-301, 5-303 | Compliance with the audit and notice procedure. 6 RCNY § 5-301 | The bias audit itself is a disparate-impact assessment, but liability attaches to the process, not to the underlying disparity. 6 RCNY § 5-301 |
| Colorado | AI Act (SB 24-205) Colorado Attorney General guidance·SB 24-205 | Reasonable-care duty. SB 24-205 § 6-1-1702 | Failure to use reasonable care to protect consumers from algorithmic discrimination. SB 24-205 § 6-1-1702 | Rebuttable presumption defense for compliance with the risk-management program and impact-assessment requirements. SB 24-205 §§ 6-1-1702, 6-1-1706 | Duty-of-care framework rather than direct disparate-impact liability. SB 24-205 § 6-1-1702 |
| California | Automated-Decision Systems (ADS) Regulations under the Fair Employment and Housing Act (FEHA) California Civil Rights Department announcement·Final FEHA ADS regulations | Disparate impact. 2 CCR § 11017(e) | FEHA disparate-impact framework, with the federal Uniform Guidelines on Employee Selection Procedures incorporated. 2 CCR § 11017(e) | Job-relatedness + business necessity (traditional FEHA framework). 2 CCR § 11017 | The vendor (as the employer's agent) is also liable. 2 CCR § 11008(b) |
| Texas | TRAIGA (HB 149) HB 149 bill page·HB 149 text | Intent-based. HB 149 § 552.052 | Developing or deploying an AI system "with the intent to" discriminate against a class protected by federal or state law. Disparate impact alone is not a violation. HB 149 § 552.052 | 60-day cure period; rebuttable-presumption defense for substantial compliance with the National Institute of Standards and Technology AI Risk Management Framework (NIST AI RMF) or a comparable framework. HB 149 §§ 552.106, 552.108 | Applies broadly to any developer or deployer doing business in Texas (not just government), but the discrimination prohibition is intent-based only. HB 149 §§ 552.002, 552.052 |
| Jurisdiction | Law | Enforcer | Private right of action | Damages / penalties | Cure period |
|---|---|---|---|---|---|
| Illinois | HB 3773 Illinois Department of Human Rights guidance·Public Act 103-0804·Draft IDHR Subpart J Rules | Illinois Department of Human Rights (IDHR) → Illinois Human Rights Commission (IHRC) → Circuit Court, jury available on election. IHRA Articles 7A & 8 | Yes, via IHRA exhaustion to circuit court. 775 ILCS 5/7A-102 | Uncapped emotional distress + back pay + attorneys' fees + injunctive relief; civil penalties of $16,000–$70,000 under SB 2487. 775 ILCS 5/8A-104·IDHR guidance | None. PA 103-0804 (no cure period) |
| New York City | Local Law 144 Department of Consumer and Worker Protection guidance·6 RCNY § 5-300·NYC Admin Code Subchapter 25 | NYC Department of Consumer and Worker Protection (DCWP). NYC Admin Code § 20-872 | No. NYC Admin Code § 20-872 | $500 to $1,500 per day; weak enforcement track record per the December 2025 NYC Comptroller audit. NYC Admin Code § 20-872(a) | None specified. NYC Admin Code § 20-872 |
| Colorado | AI Act (SB 24-205) Colorado Attorney General guidance·SB 24-205 | Colorado Attorney General only. SB 24-205 § 6-1-1706 | No. SB 24-205 § 6-1-1706 | Civil enforcement by the Colorado Attorney General. SB 24-205 § 6-1-1706 | None specified. SB 24-205 § 6-1-1706 |
| California | Automated-Decision Systems (ADS) Regulations under the Fair Employment and Housing Act (FEHA) California Civil Rights Department announcement·Final FEHA ADS regulations | California Civil Rights Department (CRD), plus a private FEHA suit. FEHA ADS regulations | Yes, via the FEHA private right of action. Cal. Gov. Code § 12965 | FEHA standard damages. Cal. Gov. Code § 12965 | None. FEHA ADS regulations (no cure period) |
| Texas | TRAIGA (HB 149) HB 149 bill page·HB 149 text | Texas Attorney General only. HB 149 § 552.107 | No. HB 149 § 552.107 | $10,000 to $200,000 civil penalties. HB 149 § 552.107 | 60 days. HB 149 § 552.108 |
Illinois HB 3773
Illinois HB 3773 bars employers from using AI in ways that discriminate, requires notice when AI is used in employment decisions, and defines AI broadly to cover most HR tools.
Effective date
January 1, 2026.
Draft rules released late 2025. Final rules pending.
Covered employers
- Any Illinois employer with 1 or more employees for 20+ calendar weeks.
- For disability, pregnancy, and sexual-harassment claims: 1 or more employees, no week threshold.
- Under the draft rule: employer "agents," including recruiters and third-party vendors.
Covered decisions
Recruitment, hiring, promotion, renewal, selection for training or apprenticeship, discharge, discipline, tenure, and any terms, privileges, or conditions of employment.
AI definition
Broad: covers any software that, when prompted, returns text, image, audio, or video content.
Compliance requirements
1. Notice requirements
- Cadence: annual for existing employees; 30 days when new tools are adopted.
- Channels: handbook or policy doc; physical premises; intranet or external website; job postings.
- Required content (7 items): product name and vendor; covered decision; AI purpose and data categories; positions affected; HR contact; accommodation rights; the Illinois Human Rights Act (IHRA) § 2-102(L) anti-discrimination language.
- Retention period: 4 years.
2. Audit tools
Use of AI tools must not result in disparate impact. Intent not required. Bans use of ZIP codes as a proxy for protected classes.