Deployment Labs
Deployment NotesAI regulation

Illinois AI Law: What Employers Must Do Under HB 3773

Deployment LabsUpdated 12 min read

Illinois has had an AI law for employers since January 1, 2026, and most of what's written about it describes rules that don't exist. The law requires notice whenever AI is used in a job decision, and it left the details of that notice to a state agency. That agency proposed them in May 2026 and withdrew them three weeks later. The duty didn't go away. It just has no instructions. Meanwhile, the Illinois AI law that made headlines in July 2026 is about companies that train frontier models, not the ones that use AI to screen résumés.

What is the Illinois AI law?

The Illinois AI law for employers is House Bill 3773 (Public Act 103-0804), an amendment to the Illinois Human Rights Act in force since January 1, 2026: it makes it a civil rights violation to use AI that has the effect of discriminating against protected classes in any employment decision, to use zip codes as a proxy for a protected class, or to fail to tell employees that AI is being used for those decisions.

This guide is for the people who own hiring and HR tools at a company with staff or applicants in Illinois: HR, recruiting, operations and IT leads. It covers HB 3773, the separate video interview law, and where the 2026 frontier AI law fits. It's based on the statutes as published by the Illinois General Assembly. It isn't legal advice.

Which Illinois AI laws apply to your company?

Illinois has three AI laws a company might be asked about, and for most employers only the first two matter. The table shows who each one reaches.

Law Who it reaches In force What it asks
HB 3773 (Human Rights Act, Section 2-102(L)) Employers with staff in Illinois January 1, 2026 No AI with a discriminatory effect, no zip code proxies, notice to employees
Artificial Intelligence Video Interview Act Employers using AI to analyze recorded video interviews for Illinois-based jobs January 1, 2020 Notice, explanation and consent before the interview; a yearly report in some cases
AI Safety Measures Act (SB 315) Developers of frontier AI models January 1, 2027 Safety frameworks, transparency reports, incident reporting, audits

Who does HB 3773 apply to?

HB 3773 applies to any employer covered by the Illinois Human Rights Act, which for most claims means anyone employing one or more employees within Illinois during 20 or more calendar weeks in the current or preceding calendar year. For claims based on disability, pregnancy or sexual harassment, one employee is enough with no week count. The state, public bodies and parties to public contracts are covered regardless of size.

The law's definition of AI is broad: "a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments." It adds that AI "includes generative artificial intelligence." So a résumé ranker counts, and so does a manager asking a chatbot to draft a performance warning.

The decisions it covers are the whole employment life cycle: "recruitment, hiring, promotion, renewal of employment, selection for training or apprenticeship, discharge, discipline, tenure, or the terms, privileges, or conditions of employment."

Fig. 1What HB 3773 asks once AI is used in any covered employment decision. Training means selection for training or apprenticeship. Source: 775 ILCS 5/2-102(L).

What does HB 3773 prohibit?

HB 3773 bans two uses of AI in employment decisions: AI "that has the effect of subjecting employees to discrimination" on the basis of a protected class, and using "zip codes as a proxy for protected classes."

Effect, not intent

The key word is "effect." Illinois doesn't need to show you meant to discriminate. If a screening tool rejects older applicants, or women, at a higher rate, and that can't be justified, the tool's effect is the violation. That's the opposite of the Texas AI law, which says "a disparate impact is not sufficient by itself to demonstrate an intent to discriminate."

The law doesn't require a bias audit, and it doesn't say how to measure effect. But testing a tool's outcomes by protected class is the only way to find out whether you're exposed before a charge tells you.

Zip codes

The zip code ban targets a familiar shortcut. A model that scores applicants on commute distance or home address can end up sorting them by race, because neighborhoods are segregated. Illinois named it directly. Check whether any tool you use takes an address, zip code or distance as an input, and ask the vendor what it does with it.

What notice does the Illinois AI law require?

The statute makes it a violation "to fail to provide notice to an employee that the employer is using artificial intelligence" for the covered decisions, and it leaves the details to the Illinois Department of Human Rights. The law tells the department to adopt rules on "the circumstances and conditions that require notice, the time period for providing notice, and the means for providing notice."

The rules were proposed, then withdrawn

The department published proposed notice rules in the Illinois Register on May 15, 2026. On June 2, 2026, it said it was withdrawing and postponing them, and postponed the public hearing set for June 10. Its stated reason, as reported by the law firm Ogletree Deakins: "Postponement is necessary to allow for continued collaboration with other state agencies." As of November 2, 2026, we haven't found a refiled version, and the department hasn't given a new timeline.

None of that suspends the statute. The notice duty has applied since January 1, 2026, with no rule saying exactly how to meet it.

What the withdrawn draft would have required

A withdrawn draft isn't binding. It is the best evidence of what the department expects, and building your notice to its standard is the safest reading of a vague duty. According to a summary by the law firm Kilpatrick Townsend, the draft would have required notice to "employees, applicants, and exclusive bargaining representatives," given:

  • When: within 30 days of adopting a new or substantially updated AI system, and in job postings for applicants;
  • Where: in employee handbooks or policy documents, at worksites where notices are usually displayed, and on the employer's intranet and external website;
  • How: in plain language, in the languages commonly spoken by the workforce, and accessible to people with disabilities.

It listed six things each notice had to say.

Fig. 2What the Department of Human Rights' withdrawn May 2026 draft would have required in every AI notice. It isn't binding, but it's the best guide to what the department expects. Source: Kilpatrick Townsend's summary of the draft (secondary).

The draft would also have extended the retention period for employment records from one year to three, and required employers to keep records of their AI notices.

Do applicants get notice?

The statute says notice goes to "an employee," and the Human Rights Act defines employees as people performing paid services in Illinois, apprentices and applicants for apprenticeships. Yet the ban covers recruitment and hiring, where the people affected are applicants, and the withdrawn draft named applicants explicitly. Until the rules return, put the notice in your job postings and on your careers page too. It costs little and closes the gap.

What does the AI Video Interview Act require?

Since January 1, 2020, an employer that asks applicants for recorded video interviews and uses AI to analyze them must, before the interview, tell each applicant AI may be used, explain "how the artificial intelligence works and what general types of characteristics it uses to evaluate applicants," and get their consent. It applies when considering applicants "for positions based in Illinois." An employer "may not use artificial intelligence to evaluate applicants who have not consented."

Three more rules follow:

  • Sharing. You may share applicant videos only with people "whose expertise or technology is necessary" to evaluate the applicant.
  • Deletion. If an applicant asks, you must delete their interviews within 30 days and tell anyone you shared them with to delete theirs, including backups.
  • Reporting. If you rely "solely upon an artificial intelligence analysis" to decide who gets an in-person interview, you must collect the race and ethnicity of applicants who did and didn't get one, and of those hired, and report it to the Department of Commerce and Economic Opportunity every year by December 31.
Fig. 3The yearly cycle for employers that rely solely on AI analysis of video interviews to choose who gets an in-person interview. The report goes to the Department of Commerce and Economic Opportunity. Source: 820 ILCS 42/20.

The report covers the 12 months ending November 30. If your hiring flow uses AI-scored video as the only gate to an in-person interview, the next report is due December 31, 2026, covering December 1, 2025, to November 30, 2026. The duty sits with the employer, not the vendor, so check with whoever bought the video tool.

How is the Illinois AI law enforced?

HB 3773 has no penalty of its own: a violation is a civil rights violation under the Human Rights Act, enforced through the same charge process as any discrimination claim. An applicant or employee can file a charge with the Department of Human Rights within 2 years of the violation.

If the Human Rights Commission finds a violation, it can order the employer to:

  • stop the practice;
  • pay actual damages;
  • hire, reinstate or promote the complainant, with back pay;
  • pay the complainant's attorney and expert witness fees;
  • pay a civil penalty for each violation and each aggrieved party, of up to $16,000 for an employer with no prior violation, up to $42,500 with one in the past five years, and up to $70,000 with two or more in the past seven years.

Because penalties can apply per aggrieved party, a tool that screens thousands of applicants carries more risk than one used on a few promotions.

What about the AI Safety Measures Act?

The Illinois law signed on July 6, 2026 (Senate Bill 315, the Artificial Intelligence Safety Measures Act) applies to developers of frontier AI models, not to companies that use AI. It takes effect January 1, 2027.

A frontier model, under the act, is a foundation model trained with "a quantity of computing power greater than 10^26 integer or floating-point operations." The heaviest duties fall on large frontier developers, those with annual revenue, together with affiliates, "in excess of $500,000,000." They must write, publish and follow a frontier AI framework and submit to annual independent third-party audits. The act also requires transparency reports before new frontier models are deployed, and reports of critical safety incidents. It creates no private right of action.

If your company buys ChatGPT, Copilot or an AI hiring tool, this law doesn't put new duties on you. It's the reason "Illinois AI law" searches spiked in July 2026, and the reason some summaries mix it up with HB 3773.

Notice is an inventory problem

The notice duty sounds like a writing task. It's really a question of knowing what you run. You can't tell employees that AI is used in hiring, discipline or scheduling if nobody can say which tools do it, and in 2026 that list includes features vendors switched on inside software you already pay for: an applicant tracking system that now ranks candidates, a performance platform that drafts reviews.

That's why our framework starts guardrails with a list of what's running and an owner for each item. In our free course, Your AI list builds the register, and Owners and backups puts a name against each tool, so that someone is responsible for the notice when a tool changes. With those two in place, the notice almost writes itself: it's the register, filtered to employment decisions and translated into plain language.

How to comply with the Illinois AI law

There are six steps. Each produces something the next one uses.

01 · List every AI tool that touches an employment decision

Cover all of them: sourcing and ad targeting, résumé screening, assessments, video interviews, scheduling, performance reviews, discipline and promotion. Include generative AI that managers use to draft decisions. The free Shadow AI Audit finds the AI features inside existing software. For each tool, record the vendor, the decisions it affects, the jobs it's used for and its owner.

02 · Ask each vendor three questions

Does the tool use address, zip code or distance as an input? What testing has the vendor done for adverse impact by protected class, and can you see the results? What data does it collect about candidates and employees? Keep the answers with your register. They feed step 03 and step 04.

03 · Write one notice, to the withdrawn draft's standard

Cover the six items in Fig. 2 for each tool: developer and product, the decisions, the purpose and data, the jobs, a contact person, and how to request an accommodation. Write it in plain language and in the languages your workforce speaks. Publish it in the handbook, on the intranet, on your careers page and in job postings, and update it within 30 days of adding or substantially changing a tool. Your AI usage policy should point to it.

04 · Test outcomes before a charge does

For high-volume tools, compare selection rates by sex, race and age, using your own data or the vendor's. Where one group fares markedly worse, find out why before you keep using the tool. Remember that Illinois looks at effect, so "we didn't mean it" isn't a defense.

05 · Check the video interview workflow

If any role uses recorded video interviews scored by AI, confirm the notice, explanation and consent happen before the interview, that videos are shared only as the law allows, and that deletion requests are honored within 30 days. If AI alone decides who gets an in-person interview, collect the race and ethnicity data and file the report by December 31.

06 · Keep records and set a review date

Keep copies of every notice version, when and where it was posted, and your test results. Set a date to review the register and notice, and check for refiled rules then. For a company that wants the register, notices and testing set up across Illinois and other states with AI hiring rules, such as Colorado and Connecticut, that's what our AI governance work does.

What's still open

  • The notice rules. The Department of Human Rights still has to adopt rules on when, how and in what form notice is given. When it does, they may differ from the withdrawn draft.
  • Applicants. Whether the statutory notice reaches applicants, not just employees, is unsettled until the rules or a decision say so.
  • How "effect" is measured. The law sets no test or threshold for a discriminatory effect. The first charges and Commission decisions under Section 2-102(L) will show how it's applied.
  • Federal pressure. Executive Order 14365, signed December 11, 2025, told the US attorney general to set up an AI Litigation Task Force within 30 days whose "sole responsibility" is to challenge state AI laws that conflict with federal policy. As of this update, HB 3773 remains in force.

We'll update this article when the department refiles its rules or the first decision under the law is published. The date at the top of the page shows when we last checked.

Sources

Questions

Common questions

Start with step one

Find every AI tool that touches a job decision.

You can't give notice about AI you don't know you use. The free Shadow AI Audit lists the AI tools in use across your company in about five minutes, including the AI features vendors have switched on inside your applicant tracking and HR software.

Run the Shadow AI AuditSee our AI governance work