Compliance · July 23, 2026 · 10 min read

Illinois AI Hiring Law: AIVIA, HB 3773, and How to Comply

Illinois AI hiring law explained: the AI Video Interview Act plus HB 3773's 2026 Human Rights Act amendments, and how to run a compliant AI assessment process.

← Part of Compliance-First Hiring AI: LL144 and the EU AI Act

Illinois AI hiring law is no longer one narrow statute about video interviews. As of January 1, 2026, it is a layered regime: the AI Video Interview Act (AIVIA) governs how you collect and analyze interview video, and HB 3773 amends the Illinois Human Rights Act (IHRA) to make discriminatory AI use in any employment decision a civil rights violation — with a notice requirement attached. If you screen, assess, or interview candidates in Illinois, both layers now apply to you at once.

This guide is for talent leaders, recruiters, and compliance owners at companies that hire in Illinois or run multistate funnels that include Illinois candidates. You will get a plain-language breakdown of what each law actually requires, what changed on January 1, 2026, where the rulemaking currently stands as of mid-2026, how Illinois compares with NYC Local Law 144, and a concrete blueprint for an assessment process that satisfies both statutes.

The timing matters. In 2026, AI sits on both sides of the hiring table: candidates use it to polish applications and, sometimes, to cheat assessments, while employers use it to generate questions, score responses, and rank applicants. Illinois' answer is not to ban the technology — it is to make employers own the outcomes. That shifts the compliance question from "do we use AI?" to "can we show our AI-assisted process does not produce discriminatory effects?" That is a question about evidence, and most hiring stacks were not built to produce it.

Why does Illinois matter so much in AI hiring regulation?

Illinois has a habit of regulating early and litigating hard. Its Biometric Information Privacy Act spawned some of the largest privacy settlements in US history, and AIVIA — in force since January 1, 2020 — was the first US law specifically targeting AI analysis of job interviews. When Illinois attaches a new obligation to the Human Rights Act, as HB 3773 does, it plugs AI directly into a mature enforcement machine: the Illinois Department of Human Rights (IDHR) charge process, the Human Rights Commission, and a plaintiffs' bar that knows how to use both.

That is why the Illinois AI hiring law deserves more attention than its modest text suggests. Unlike Colorado's AI Act, which builds an elaborate risk-management framework, Illinois keeps the statute short and lets the discrimination standard do the work. There is no safe harbor for "we used a reputable vendor." If the tool's output has a discriminatory effect, the employer using it is exposed.

What does the AI Video Interview Act require?

AIVIA (820 ILCS 42) applies when an employer asks applicants to record video interviews for Illinois-based positions and uses AI to analyze them — evaluating word choice, facial expression, or any other characteristic to judge fitness for the role. Its obligations fall into four buckets.

Notice, explanation, and consent

• Notify each applicant before the interview that AI may be used to analyze their video and evaluate their fitness for the position.
• Provide an explanation of how the AI works and the general types of characteristics it uses to evaluate applicants — a real description, not a vague disclaimer.
• Obtain consent before the interview. Recent amendments tightened this to explicit written consent: proceeding with the interview no longer counts as agreement.
• If a candidate declines, you cannot analyze their video with AI — which means you need a genuinely equivalent non-AI evaluation path.

Sharing limits and deletion

Employers may share interview videos only with people whose expertise or technology is necessary to evaluate the candidate. On request, the employer must delete all copies within 30 days — and instruct anyone else holding copies, including backups, to do the same. That reaches into your vendor contracts: if your interview platform cannot execute a cascading deletion, you cannot comply.

Demographic reporting

Since 2022, employers that rely solely on AI analysis to decide who advances to an in-person interview must collect the race and ethnicity of applicants and report that data to the state annually. The legislative intent is unambiguous: Illinois wants a statistical record from which algorithmic bias can be detected. If your funnel has a stage where AI alone gates advancement, you have a reporting obligation — and a strong incentive to keep a human decision in that loop.

What does HB 3773 change on January 1, 2026?

HB 3773, signed in August 2024 and effective January 1, 2026, amends the Illinois Human Rights Act rather than creating a standalone AI statute. That design choice is the whole story: AI discrimination is now handled like any other civil rights violation, with the same remedies and the same enforcement path. Three provisions matter most.

No discriminatory effect

Employers may not use AI that has the effect of subjecting employees or applicants to discrimination on the basis of protected classes — across recruitment, hiring, promotion, renewal, selection for training or apprenticeship, discharge, discipline, tenure, and the terms, privileges, or conditions of employment. "Effect" is the operative word. Intent is irrelevant; a neutral-seeming tool that screens out a protected group disproportionately is a violation. That makes ongoing adverse-impact analysis the practical center of compliance: you cannot prove the absence of a discriminatory effect without measuring selection rates by group.

No zip-code proxies

The amendment expressly prohibits using zip code as a proxy for protected classes. This is aimed at a well-documented machine-learning failure mode: models trained on historical data learn that geography predicts outcomes because geography correlates with race and national origin. Illinois calls the technique out by name, and any assessment vendor whose scoring pipeline ingests location data should expect hard questions about it.

Notice — and a rulemaking still in motion

Employers must notify employees and applicants when AI is used for the covered employment decisions. What that notice must say, and when it must be delivered, was supposed to be settled by IDHR regulation. The department published proposed notice rules on May 15, 2026, then withdrew them on June 2, 2026, citing the need to coordinate with other state agencies. As of July 2026, the statutory notice duty is fully in force but the implementing details are pending. The defensible posture in the meantime: give clear, specific, pre-use notice that names the decision stage, describes the AI's role in it, and explains what the tool evaluates — the same substantive content the draft rules contemplated.

How does Illinois compare with NYC Local Law 144?

Most multistate employers already know NYC Local Law 144, so the comparison is a useful anchor. LL144 is audit-first: before using an automated employment decision tool, you need an independent bias audit within the past year, a public summary of results, and advance notice to candidates. Illinois is effect-first: no audit mandate, no publication requirement — but a flat prohibition on discriminatory effects, enforced through a civil rights statute with real remedies.

• Scope: LL144 covers a defined class of automated employment decision tools; Illinois' HB 3773 covers essentially any AI that influences an employment decision, with no carve-outs.
• Obligation: LL144 requires you to test and publish; Illinois requires the outcome itself to be non-discriminatory, tested or not.
• Notice: both require candidate notice; Illinois' precise format awaits final IDHR rules, while NYC's is specified.
• Enforcement: LL144 is enforced by a city agency with per-violation fines; Illinois routes through IDHR charges, with damages, penalties, and attorneys' fees on the table.
• Practical convergence: under either regime, the winning strategy is identical — structured, job-related assessments plus routine adverse-impact monitoring and documented human review.

The convergence point is worth underlining. An employer that runs bias audits for NYC already generates most of the evidence needed to defend an Illinois charge; an employer that monitors adverse impact for Illinois is most of the way to an LL144 audit. This is the core argument of compliance-first hiring: build one evidence-producing process and satisfy many regimes, rather than bolting a checklist onto each jurisdiction.

What does a compliant AI assessment process look like in Illinois?

Illinois' layered rules reward a specific process shape: job-related by construction, transparent to candidates, human-decided, and instrumented for evidence. Here is the pipeline we recommend, stage by stage.

• Inventory and classify. List every tool in your funnel that generates outputs influencing decisions — resume screeners, assessment scorers, interview analyzers, ranking features inside your ATS. Under HB 3773's broad AI definition, more of your stack is covered than you think.
• Anchor assessments to the job. Generate or select assessment content from the actual job description, so every question maps to a stated requirement. Job-relatedness is your first line of defense against any discriminatory-effect claim, and it starts with writing the job description properly.
• Notify before use, in plain language. Tell candidates which stages use AI, what it evaluates, and what role it plays in the decision. For video interviews, add AIVIA's explanation-and-written-consent step and a real alternative path for candidates who decline.
• Keep a human decision in the loop. Avoid any stage where AI alone gates advancement — it triggers AIVIA's demographic-reporting duty and concentrates discriminatory-effect risk in an unreviewed step.
• Score against structured rubrics. Predefined, job-related criteria applied uniformly — the logic of structured interviews extended to every assessment stage — make outcomes explainable and comparable.
• Monitor adverse impact continuously. Compute selection rates by group at each funnel stage, not once a year. A quarterly four-fifths check catches drift before it becomes a charge.
• Operationalize deletion and retention. Wire AIVIA's 30-day cascading deletion into your systems and contracts, and document retention periods for every artifact the process produces.

Notice what this pipeline is not: it is not a ban on AI, and it is not a return to static tests. Illinois is agnostic about the technology and strict about the outcome. A well-instrumented AI-assisted process — per-job content, structured scoring, human decisions, continuous monitoring — is easier to defend than a legacy process built on shared question banks and unstructured interviewer judgement, because it produces the evidence a regulator will ask for as a byproduct of running.

Where do employers get Illinois AI compliance wrong?

We see the same failure modes repeatedly, and they are worth naming plainly.

• Assuming vendor compliance transfers. It does not. HB 3773 attaches to the employer using the tool; a vendor's marketing page about fairness is not evidence of non-discriminatory effect in your applicant pool.
• Treating AIVIA consent as a checkbox. Burying AI disclosure in a general terms-of-service acceptance fails the explanation requirement and, after the written-consent amendments, fails the consent requirement too.
• Waiting for final IDHR rules to give notice. The statutory notice duty took effect January 1, 2026, regardless of the withdrawn regulations. Silence while rulemaking is pending is the riskiest available position.
• Testing for bias once, at procurement. Model behavior and applicant pools both drift. A point-in-time evaluation from 2024 says little about your 2026 selection rates.
• Ignoring the candidate-side arms race. If AI-assisted cheating corrupts your assessment scores, it corrupts your adverse-impact statistics too — integrity and compliance are the same problem wearing different hats.

How should multistate employers handle Illinois?

If you hire across jurisdictions, resist the urge to build an Illinois-only workflow. The pattern across NYC LL144, Colorado's AI Act, Illinois' HB 3773, and the EU AI Act is convergent: disclose AI use, keep humans deciding, prevent and measure discriminatory outcomes, and retain evidence. Applying the strictest applicable standard funnel-wide — explicit consent for AI-analyzed interviews, universal pre-use notice, continuous impact monitoring — costs marginally more than segmenting candidates by state and eliminates an entire class of misclassification risk, such as an Illinois-based remote candidate slipping through a workflow keyed to office location.

It also produces a better candidate experience. Candidates increasingly expect to be told when AI evaluates them; employers that disclose clearly and offer alternatives read as confident, not evasive. Compliance and employer brand point the same direction here.

Where H-Evaluate fits

H-Evaluate was designed compliance-first for exactly this regulatory shape. Assessments are generated per job description with quality-gated generation — no static shared test library — so every question is traceable to a stated job requirement, which is the foundation of any job-relatedness defense. Sandbox work samples put every candidate through the same task under the same conditions, and structured scoring produces the explainable, comparable records that both AIVIA transparency and HB 3773's discriminatory-effect standard demand. Human reviewers make the advancement decisions; the system's job is to give them better evidence.

None of this makes Illinois compliance automatic — no tool can, and vendors claiming otherwise are selling you exposure. What an AI-native hiring stack can do is make the compliant process the default process: notice built into the candidate flow, job-derived content by construction, and audit-ready artifacts as a byproduct of every assessment rather than a scramble after a charge is filed.

Illinois didn't ban AI in hiring. It banned unaccountable AI in hiring — and the employers who thrive under HB 3773 will be the ones whose processes were built to produce evidence, not excuses.