
Short answer: as of 12 October 2026, five places have AI hiring rules you must follow today: New York City, Illinois, California, Texas and Connecticut. Colorado's law and Connecticut's employer notice duties are passed but not yet in force. Most of these rules ask for notice and no bias. Of the laws we read, only New York City asks for a yearly bias audit, an outside test of whether a tool treats groups differently. An automated tool here means software that scores, ranks or sorts job candidates.
Written and checked on 12 October 2026. This is general information, not legal advice; check the cited statute or opinion before relying on it.
Official text, as of 12 October 2026
Illinois Human Rights Act, 775 ILCS 5/2-102(L)(2), added by Public Act 103-0804 (HB 3773). It is a civil rights violation "for an employer to fail to provide notice to an employee that the employer is using artificial intelligence for the purposes described in paragraph (1)." Section 99 of the act says it "takes effect January 1, 2026." See the section on the General Assembly site and the enrolled act.
What changed in 2025 and 2026: California's rules began on 1 October 2025. Illinois and Texas began on 1 January 2026. Connecticut's first hiring rule began on 1 October 2026. Colorado replaced its law and set a new start of 1 January 2027.
TL;DR
- In force now: New York City, Illinois, California, Texas, and Connecticut's "AI is no defense" rule.
- Passed, not in force: Colorado (1 January 2027) and Connecticut's notice duties (tools used on or after 1 October 2027).
- Most duties are notice and no bias, and none asks you to get an agency's approval first.
- Federal help moved, since the EEOC's two AI pages now show "not found."
- Still moving: the Illinois notice rules and the Colorado rules are unfinished. Recheck before you rely on this page.
In force today
| Law | Covers | What it asks | Since |
|---|---|---|---|
| NYC Local Law 144 of 2021 (DCWP page) | Hiring and promotion in the city | A bias audit within one year before use, a public summary, and notice 10 business days ahead | Enforced from 5 July 2023 |
| Illinois Human Rights Act, 775 ILCS 5/2-102(L) | Hiring, promotion, discipline, discharge and more | No AI with the effect of discrimination, no zip codes as a stand-in for a protected group, and notice to employees | 1 January 2026 |
| Illinois AI Video Interview Act, 820 ILCS 42/5 | Applicants asked to record video interviews | Tell the applicant, explain how the AI works, get consent first | 1 January 2020 |
| California, 2 CCR 11009(f) and 11013(c) | Employers and other covered entities | Discriminatory automated tools are unlawful, and tool data is kept four years | 1 October 2025 |
| Texas, Bus. & Com. Code 552.056 | Anyone who develops or deploys AI | No AI built or used with intent to discriminate unlawfully | 1 January 2026 |
| Connecticut, Public Act 26-15, sections 13 and 14 | Employers | Using an automated tool "shall not be a defense" to a bias complaint under the state's employer-discrimination law | 1 October 2026 |
A few details from the text:
- New York City: The state Comptroller says the city can fine $500 to $1,500 a day. Our NYC Local Law 144 guide has the details.
- California: Section 11008.1 calls an automated-decision system "a computational process that makes a decision or facilitates human decision making regarding an employment benefit." Section 11009(f) makes a biased one unlawful. It also says proof of bias testing, or the lack of it, matters. The Civil Rights Council announcement gives the dates. See also California's other AI laws.
- Texas: Section 552.056(c) adds that "a disparate impact is not sufficient by itself to demonstrate an intent to discriminate." The state attorney general (AG) has "exclusive authority" to enforce it. A licensing agency may add sanctions if the AG recommends them. Section 552.101(b) gives no one a right to sue. See the Texas guide and the enrolled bill.
- Connecticut: The no-defense rule is in the act signed on 27 May 2026. Section 13 adds it to General Statutes 46a-60(b)(1), the employer part of the state's bias law. Section 14 does the same in 46a-81c. The commission or a court "may consider evidence of anti-bias testing," but the act does not require testing.
Enacted, not in force yet
| Law | Starts | What it will ask |
|---|---|---|
| Colorado SB 26-189 (signed act) | Applies to decisions made on or after 1 January 2027 | Notice, then a plain explanation within 30 days of a bad outcome |
| Connecticut Public Act 26-15, sections 7 to 12 | Tools deployed on or after 1 October 2027 | Plain-language disclosure when a tool talks to applicants, and written notice before an employment decision |
| California privacy rules, 11 CCR 7200(b) | Compliance by 1 January 2027 | Rules for a business using ADMT (automated decision-making technology) for a "significant decision" |
Colorado: Hiring is a covered area, and an applicant who lives in Colorado counts as a consumer. Only the state AG can enforce the law. If the AG sees a fix as possible, a notice and 60 days to cure come first. Knowing or repeated violations get no cure period. The law gives no new right to sue. Draft rules are open for comment through 26 October 2026, with a hearing that day (coag.gov). A court order from 27 April 2026 in X.AI LLC v. Weiser (D. Colo.), which the AG agreed to, also holds back AG enforcement for now, on our reading of its wording. It does not repeal the law. Our Colorado guide goes section by section.
Connecticut: Only the AG enforces sections 8 to 11. For violations on or before 31 December 2027, the AG may send a notice and allow 60 days to fix the problem.
California privacy rules: Section 7001 counts "employment or independent contracting opportunities or compensation" as a "significant decision." The rules began on 1 January 2026. Section 7200(b) gives a business already using the tool until 1 January 2027. A written privacy risk assessment is also due before such use (section 7150(b)(3)). It looks at privacy risk, not how groups fare, so we do not count it as a bias audit.
The one sentence worth reading closely
The Illinois law is short. It bans AI that "has the effect of subjecting employees to discrimination." It bans skipping the notice. And it tells the Illinois Department of Human Rights to write rules on "the circumstances and conditions that require notice, the time period for providing notice, and the means for providing notice."
Those rules are not done. The Department proposed them in May 2026 (50 Ill. Reg. 6794) and put the work on hold on 2 June 2026. The Illinois Human Rights Commission's June 2026 training slides list both steps. The Illinois Register summary for 26 June 2026 says the Department withdrew the proposal "to further consult and collaborate with other State agencies." A Seyfarth alert says no new timeline was given. We did not find a new proposal, and the Department's own news page does not show the withdrawal, so check the Register for later issues. The duty to give notice is still in the statute. Our Illinois citation guide shows where to look it up.
What federal law says now
The EEOC once posted two guides on AI and civil rights law. As of 12 October 2026, both pages show "not found" on eeoc.gov. A K&L Gates alert reports that the EEOC took down its AI guidance on 27 January 2025. We did not verify why.
Executive Order 14281 (90 Fed. Reg. 17537) says it is US policy "to eliminate the use of disparate-impact liability in all contexts to the maximum degree possible" (Section 2). Section 4 tells agencies to "deprioritize enforcement" of laws and rules to the extent they include it. An order is not a statute. The civil rights laws still apply. We found no federal statute that sets AI hiring rules or overrides these state laws. Our AI regulation update covers the federal side, and the EEOC data guide covers the agency's documents.
What goes wrong
- Assuming one vendor audit covers every state, when only New York City asks for one, on its own timing.
- Treating "a human decides" as a free pass, although each law sets its own reach.
- Giving notice in one place but not another for the same tool.
- Using data that stands in for a protected group, such as the zip codes Illinois names.
- Relying on draft rules, since Illinois paused its rules and Colorado's are still drafts.
- Reading a 2025 explainer in late 2026, after the Colorado and Connecticut dates moved.
Questions to ask your hiring-tool vendor
- Which of our locations does the tool run in, and which laws have you mapped?
- Can you give us the audit summary New York City requires, and how recent is it?
- What notice wording can you supply for each state?
- What data does it use? Does any of it stand in for a protected group?
- How does it handle an applicant who asks for a human review?
- What do you provide for Colorado's 1 January 2027 date?
What we did not cover
This post covers six places, not every state. Other states and cities may have rules we did not read. If your state is not here, read its code before you assume it has no rule.
How we keep this current: we recheck every date and status on this page every 60 days until January 2027. The next check is due by 11 December 2026.
A 40-person Chicago company uses software that ranks resumes. Besides avoiding discriminatory effects, what does the Illinois Human Rights Act amendment require?
FAQ
Is AI hiring regulated in the US? Yes, mostly by state and city laws. We found no federal statute that sets AI hiring rules for private employers. Federal civil rights laws still apply when software does the hiring.
Which states have AI hiring laws in force? Of the places we read: Illinois, California, Texas, Connecticut (a no-defense rule) and New York City. Colorado's law starts for decisions made on or after 1 January 2027.
Do I have to tell applicants I use AI? In New York City, yes, 10 business days ahead. Illinois requires notice to employees. For video interviews it also requires notice and consent from applicants. Colorado and Connecticut add notice duties later on. Read each law's exact words.
Do I need a bias audit? New York City requires one. The other laws we read do not.
What is the Illinois AI hiring law? It is part (L) of section 2-102 of the Illinois Human Rights Act, in force since 1 January 2026. It bans AI that has the effect of discriminating and requires notice. The notice rules are still unfinished.
Are the EEOC's AI guidance pages still online? No. Both documents return "not found" as of 12 October 2026. Title VII and the Americans with Disabilities Act still apply.
How do I stay current? Recheck the dates and the official pages linked here. Our guide to tracking state law changes explains how, and the complete guide to US law data maps the sources.
For the US statutes, regulations and court rules these laws sit in, see the US primary law API.
New legal AI guides, weekly.
Further Reading
AI Chatbot Disclosure Laws by State: When a Bot Must Say It Is AI
Read postCalifornia AI Laws Explained: SB 53 and the Rest
Read postNew York RAISE Act Explained: Who It Covers and When It Starts
Read postTRAIGA: Texas AI Law Explained From the Statute
Read postUtah AI Law Explained From the Statute: What Is in Force Now
Read postNYC Local Law 144 Explained: Bias Audits for AI Hiring Tools
Read post
Co-Founder & CTO
Priyansh leads engineering and AI at Vaquill AI: the pipelines that pull statutes, regulations and court rules from every US jurisdiction's official publisher, and the REST API, MCP server and open dataset that serve them.