AI Hiring Laws Are Live in California, Illinois, and NYC — Colorado's June Deadline Makes It Urgent
AI Hiring Laws Are Live in California, Illinois, and NYC — Colorado's June Deadline Makes It Urgent
If you're an HR leader using AI tools to screen resumes, score candidates, or analyze video interviews, you're already operating under active law in three US jurisdictions. And in 93 days — on June 30, 2026 — Colorado joins them.
This isn't a far-off compliance horizon. California's rules have been in force since October 1, 2025. Illinois activated its dual AI hiring regime in January 2026. New York City has been enforcing its bias audit mandate since July 2023. The regulatory patchwork that HR teams once dismissed as hypothetical is now a real, multi-state compliance obligation — and the convergence of these frameworks at mid-2026 is creating urgent pressure on people ops and legal teams alike.
Here's what you need to know, state by state, and what to do this week.
The State-by-State Breakdown
New York City — Local Law 144 (Active since July 2023)
NYC moved first. Local Law 144 requires employers to commission annual independent bias audits of any automated employment decision tool (AEDT) used in hiring or promotion decisions within the city. Employers must post audit summaries publicly and notify candidates at least 10 days before an AI tool is used to evaluate them.
The penalty exposure is not abstract. The law imposes fines of $500 to $1,500 per violation, multiplied by the number of affected applicants and the number of days of non-compliance. Run the numbers: if 100 candidates move through a non-compliant AI screening tool over a 30-day window, that's 100 applicants × 30 days × $500–$1,500 = potential exposure of $1.5 million to $4.5 million before any corrective action credit. Miss an annual audit cycle and those figures start accumulating from day one.
California — Civil Rights Council Regulations (Active since October 1, 2025)
California's CRC employment regulations, issued under the Fair Employment and Housing Act (FEHA), apply to any employer in the state with five or more employees. The regulations take a broad view of what counts as a regulated system: resume filters, video interview analytics, aptitude assessments, and predictive hiring tools all fall within scope.
The core standard is disparate impact. Employers cannot use any automated decision system (ADS) that produces discriminatory outcomes against protected classes — regardless of intent. That's a high bar. To meet it, employers must conduct ongoing, repeatable bias testing (not just at tool launch), maintain four years of documentation covering data inputs, outputs, audit results, and decision criteria, and provide candidates with pre- and post-use notices explaining how AI was used and what appeal rights they have.
Critically, California makes clear that third-party vendors cannot shield employers from liability. You are accountable for the tools you deploy.
Illinois — Dual Regime (Active since January 2026)
Illinois operates on two parallel tracks. First, the Artificial Intelligence Video Interview Act (AIVIA) requires employers to obtain notice and consent from candidates before analyzing video interviews with AI. Second, HB 3773 amends the Human Rights Act to prohibit AI-driven employment decisions that result in discrimination based on protected characteristics — with a specific carveout banning the use of ZIP codes as proxies for protected traits.
Both tracks took effect in January 2026. Employers using AI video interviewing tools in Illinois hiring must now have notice-and-consent procedures in place. Those using any algorithmic screening system must ensure ZIP code data isn't functioning as a discriminatory proxy.
Colorado — SB 24-205 (Effective June 30, 2026)
Colorado's SB 24-205 is the countdown clock. Effective June 30, 2026, the law classifies AI systems that influence hiring, promotion, pay, or other consequential employment decisions as "high-risk." Developers and deployers of those systems must:
- Conduct and document impact assessments before deployment
- Implement transparent appeal processes so candidates can contest AI-driven decisions
- Enable affected parties to notify the state Attorney General within 90 days of a violation
Violations constitute unfair trade practices under Colorado law. The AG complaint pathway is significant for employers: candidates who experience adverse AI-driven decisions have a clear statutory route to state enforcement, which means even a single disputed AI hiring outcome could trigger regulatory attention.
Vendor Liability: The Workday Warning
AI hiring compliance isn't just an employer problem. The case Mobley v. Workday (N.D. Cal.) is testing whether HR technology vendors can be held liable as "agents" performing hiring functions on behalf of employers. The court allowed the claims to proceed and has conditionally certified an ADEA collective action — a landmark development that increases discovery burdens and settlement pressure across the industry.
The practical implication for HR buyers: your vendor contracts need to address liability allocation explicitly. Who owns the bias audit? Who bears responsibility if the tool produces discriminatory outcomes? If your current agreements don't answer these questions, they need to be updated before Colorado's deadline.
8-Step HR Compliance Checklist
Use this framework to assess your exposure and close gaps before June 30, 2026:
Inventory every AI tool in your hiring stack. Document what each tool does, which jurisdictions it's used in, and whether it influences employment decisions.
Classify high-risk tools. Under Colorado's framework, any AI system affecting hiring, promotion, or pay is high-risk. Tag these explicitly.
Commission or refresh bias audits. NYC requires annual independent audits. California and Colorado require ongoing testing. If your last audit is more than 12 months old, schedule a new one now.
Implement candidate notice and consent procedures. Required in California, Illinois (AIVIA), and NYC. Notices must explain what AI is being used and provide opt-out or appeal pathways.
Audit for ZIP code and geographic proxies. Illinois's HB 3773 explicitly bans ZIP codes as proxies for protected characteristics. Review your screening logic and vendor configurations.
Review and update vendor contracts. Post-Mobley v. Workday, liability allocation must be explicit. Ensure contracts specify audit responsibilities, indemnification scope, and data governance requirements.
Stand up 4-year record retention. California requires four years of documentation including data inputs, outputs, audit results, and decision criteria. Build this into your ATS or HRIS architecture now.
Prepare for Colorado's AG complaint pathway. SB 24-205 enables affected candidates to notify the state Attorney General within 90 days of a violation. Ensure your AI governance documentation is auditable before June 30, 2026 — an AG inquiry triggered by a candidate complaint will require you to demonstrate compliance.
What HR Teams Should Do This Week
The June 30 deadline is 93 days away (from March 29, 2026). That's enough time to act — but not enough time to delay.
Start with your inventory. Most HR leaders are surprised to discover how many AI-assisted tools touch their hiring funnel: ATS ranking algorithms, job board optimization tools, video interview platforms, and assessment scoring systems all potentially fall within scope.
Once you know what you're running, run the checklist above in order. Bias audits take weeks to commission and complete. Vendor contract renegotiations take time. The compliance infrastructure you need — documentation, notice templates, appeal processes, retention systems — doesn't build itself.
The good news: tools that are well-designed for compliance exist. Platforms like OVI that operate with human-in-the-loop architecture — where AI provides decision-support only and final hiring decisions remain with the recruiter — are structured to meaningfully reduce automated-decision exposure under frameworks like NYC's Local Law 144. Transcript-only analysis (no biometric analysis, no facial recognition, no emotion detection) further reduces AEDT classification risk. Starting at $99/month, compliant AI hiring tooling is now accessible at every budget level.
State laws are converging. The compliance crunch is here. The teams that act in the next 93 days will be in a defensible position when Colorado goes live. The ones that don't will be playing catch-up under active regulatory scrutiny.
Sources:
- Akerman — Multi-State AI Employment Law Guide: https://www.akerman.com/en/work/products/legal-alert/2025/navigating-the-multi-state-landscape-of-ai-employment-laws.html
- Harris Beach Murtha — AI Employment Law Overview: https://www.harrisbeachmurtha.com/insights/ai-in-hiring-what-employers-need-to-know/
- K&L Gates — California CRC Regulations: https://www.klgates.com/California-Civil-Rights-Council-Formally-Adopts-Employment-Regulations-for-Automated-Decision-Systems-9-30-2024
- JD Supra — Illinois AI Employment Law: https://www.jdsupra.com/legalnews/illinois-ai-employment-law-update-2026/