Quick Summary
- NYC Local Law 144 has been enforced since July 2023 and requires an independent annual bias audit, a public results summary, and 10 business days’ candidate notice before using AI in hiring with stricter enforcement expected through 2026 after a December 2025 city audit found weak follow-through.
- The EU AI Act classifies recruitment AI as “high-risk,” but its compliance deadline was formally pushed from August 2026 to December 2027 under a new law (the Digital Omnibus on AI) that entered into force in July 2026 a major, recent shift most employers haven’t fully registered yet.
- Illinois and Colorado have added their own AI hiring laws in 2026, and more states are following NYC’s model.
- None of this means AI in hiring is banned. It means employers now need documentation, audits, and human oversight built into how they use it.
If you use any AI tool to screen resumes, rank candidates, or evaluate interviews and most recruiting teams now do, often without realizing it this changed the compliance picture for 2026. Here’s exactly what’s required, what’s been delayed, and what recruiters and staffing firms need to do now.
Why This Matters Right Now
For the past two years, AI hiring tools moved faster than the laws meant to govern them. That gap is closing fast in 2026. Regulators in New York, the EU, Illinois, and Colorado have all moved from guidance to active enforcement or are about to. At the same time, one of the biggest deadlines employers had been bracing for just moved by 16 months, and most hiring teams haven’t caught up with what that actually changes.
Getting this wrong isn’t a minor compliance footnote. Penalties range from a few hundred dollars per violation in NYC to tens of millions of euros under the EU AI Act. And “our vendor handles compliance” isn’t a defense any regulator currently recognizes the obligation sits with the employer using the tool, not just the company that built it.
NYC Local Law 144 What’s Actually Required
Local Law 144 has been enforceable in New York City since July 5, 2023, and it applies to any employer or employment agency using an “Automated Employment Decision Tool” (AEDT) to screen, rank, or score candidates for NYC-based roles. That covers more than most people assume resume-scoring features in an ATS, video-interview analysis tools, and skills-assessment platforms can all qualify, depending on how heavily their output is weighted in the hiring decision.
Three things the law requires:
- An independent bias audit, every year conducted by an auditor with no financial ties to the company or the tool’s vendor, measuring selection and scoring rates across race, ethnicity, sex, and intersectional categories.
- A public summary of the audit results, posted on the employer’s website.
- Candidates notice at least 10 business days before the tool is used, candidates must be told an AEDT is in play.
| Requirement | Detail |
| Who it applies to | Employers/agencies using AEDTs for NYC-based roles |
| Audit frequency | Annual, by an independent third party |
| Candidate notice | 10 business days before use |
| First violation penalty | Up to $500 |
| Subsequent violations | $500–$1,500 each, per day of continued non-compliance |
| Enforcement status | Active since July 2023; stricter enforcement expected through 2026 |
The penalties themselves are modest compared to what’s coming out of the EU but a December 2025 audit by the NY State Comptroller found that the city agency enforcing this law had been doing it poorly, reviewing dozens of companies and flagging repeated non-compliance. Employment law firms have since warned employers to expect a noticeably stricter enforcement phase through the rest of 2026, with more proactive investigations rather than complaint-driven ones. If your bias audits, public disclosures, or candidate notices have gone stale, this is the year that catches up with you.
The EU AI Act What Changed (and What Didn’t)
This is the part most recruiters haven’t caught up with yet: the EU AI Act’s biggest deadline for employment AI just moved.
The Act classifies AI systems used for recruitment, candidate screening, ranking, interview analysis, promotion, and termination decisions as “high-risk” under Annex III. The original compliance deadline for these obligations risk management systems, documentation, human oversight, and transparency was August 2, 2026.
That deadline has now been formally pushed back. In July 2026, a new regulation called the Digital Omnibus on AI entered into force, delaying the high-risk employment obligations to December 2, 2027. This isn’t a rumor or a proposal still in negotiation; it was adopted by the European Parliament and Council and is now legally binding.
What this means practically:
- If your company (or a client) uses AI for hiring decisions affecting EU-based candidates, you now have until December 2027, not August 2026, to meet the full high-risk obligations (risk management, documentation, human oversight, bias testing).
- This is not a reason to stop preparing. The scope of what counts as “high-risk” hasn’t changed, only the timeline has. Companies that pause their compliance work now will be scrambling again in late 2027.
- One thing the delay did not touch: transparency obligations under Article 50, which already require disclosing when candidates are interacting with AI. Those apply regardless of the high-risk delay.
- Practices banned outright since February 2025 including emotion-recognition AI in the workplace remain banned. The delay only affects the high-risk category, not the prohibited one.
For staffing firms and companies hiring across borders, the practical takeaway is this: the EU AI Act still applies to you if your hiring decisions touch EU-based candidates, even if your company isn’t based in Europe. The deadline moved. The obligation didn’t disappear.
It’s Not Just NYC and the EU
The regulatory pattern NYC started in 2023 is spreading fast across the US:
- Illinois (HB 3773) : in effect since January 1, 2026, amending the Illinois Human Rights Act to prohibit AI use that produces discriminatory hiring effects, with mandatory candidate notice.
- Colorado (SB 26-189) : replaced an earlier 2024 statute in 2026 and took effect January 1, 2027, covering “high-risk AI systems” used in employment and other consequential decisions, requiring impact assessments and a formal risk management program.
More states are expected to follow this pattern through 2027. If your company operates in multiple states, treating this as “just a New York problem” is exactly the assumption that creates exposure.
Where Companies Get This Wrong
The most common mistake isn’t ignoring these laws outright; it’s assuming a vendor’s AI tool is “already compliant” because it’s marketed that way. Under both NYC Local Law 144 and the EU AI Act, the compliance obligation sits with the employer deploying the tool, not the company that built it. A vendor completing its own bias audit doesn’t automatically satisfy your legal obligation to audit the tool as you’re using it, on your candidate pool.
The second common mistake is treating the EU AI Act’s delay as permission to deprioritize the work entirely. The obligations didn’t shrink and the runway got longer. Companies that use the extra time to properly inventory their AI hiring tools and build documentation now will be in a far better position than those that wait until 2027 to start.
Our earlier piece on The Hidden Risks of AI Recruitment: What HR Leaders Must Know in 2026 covers the broader risk landscape beyond just these specific laws. And if your team is still figuring out where the ethical and legal lines sit when using AI in hiring, How to Use AI Ethically in Your Recruitment Process breaks that down in practical terms worth reading alongside this one if you’re auditing your current hiring stack.
Quick Compliance Checklist
Before your next hiring cycle, run through this:
- List every tool in your hiring stack that scores, ranks, or shortlists candidates
- Confirm whether any of those tools qualify as an AEDT under NYC Local Law 144 (if you hire for NYC-based roles)
- Check whether any hiring decisions touch EU-based candidates if yes, track your Annex III obligations against the new December 2027 deadline
- Request bias audit documentation from every AI vendor you use don’t assume it exists
- Build a candidate notice step into your process, regardless of jurisdiction
- Set a calendar reminder to re-check this in 6 months this area is moving fast
What Recruiters and Staffing Firms Should Do Now
- Inventory every tool in your hiring stack that scores, ranks, or screens candidates including features built into your ATS that you may not think of as “AI.”
- Confirm which laws apply based on where your candidates are located, not just where your company is headquartered.
- Don’t rely on vendor claims alone request documentation of any bias audits they’ve conducted, but plan for your own audit obligations regardless.
- Build candidate notice into your process now, even where deadlines have moved it’s good practice and increasingly expected by candidates themselves.
- Revisit this in six months. This regulatory landscape is moving quickly; a compliance position that’s accurate today may not be accurate by early 2027.
If you’re still mapping out exactly where AI fits into your hiring process, our breakdown of How AI Is Transforming Recruitment in 2026 is a useful starting point before you audit for compliance.
If your team is using AI as part of a broader recruitment process whether in-house or through an RPO (Recruitment Process Outsourcing) partner. This is also a good moment to confirm exactly where in that process automated tools are making or influencing decisions, since that’s the exact trigger point these laws are built around.
Final Thoughts
AI hiring tools aren’t going away, and neither is the regulation catching up to them. The practical reality for 2026 is a mixed picture: NYC’s law is entering a stricter enforcement phase, the EU’s biggest deadline just moved by 16 months, and several US states are actively adding their own versions of the same rules. None of this requires employers to stop using AI in recruitment, it requires them to be able to show their work: what tools they use, how those tools were audited, and how candidates were informed.
Companies that treat this as ongoing infrastructure not a one-time checklist will be the ones who aren’t scrambling when the next deadline lands.
Need Help Making Sure Your Hiring Process Is Actually Compliant?
Navigating AI hiring compliance alongside recruitment volume is a lot to manage internally. Book a free 15-minute consultation with Get Workz and we’ll help you understand where your current hiring process stands and what it actually needs.
FAQs About NYC Local Law 144 and the EU AI Act for Recruiters
1. Does NYC Local Law 144 apply to every AI tool recruiters use?
No. It applies to tools that meet the law’s definition of an automated employment decision tool, such as a system that substantially helps decide which candidates move forward. A tool’s name or “AI powered” label alone does not settle the question; how your team uses it matters.
2. What do recruiters need to do before using a covered tool in NYC?
The tool needs a bias audit conducted within the past year, and a summary of the results must be publicly available. Candidates also need notice at least 10 business days before the tool is used. These are steps to check before adding the tool to a hiring process.
3. Does the EU AI Act ban AI in hiring?
No. It treats many AI systems used in recruitment and employment as high risk, which means they face stricter requirements when the relevant rules take effect. Recruiters can still use AI, but they should understand what each tool does, where it influences a decision, and what safeguards are in place.
4. When do the EU rules for high risk hiring AI take effect?
Under the revised timeline, the EU AI Act’s requirements for high risk systems listed in Annex III, which include certain hiring uses, are due to apply on December 2, 2027. That gives teams time to prepare, but it does not mean every AI related rule is on hold until then.
5. If our AI vendor says its tool is compliant, is that enough?
It is a useful starting point, but ask for the evidence behind the claim. Which product features were assessed? Is the audit current? Does it cover the way your team actually uses the tool? Employers and recruiters still need to check the obligations that apply to their own hiring process.
6. Where should a recruiting team start?
Make a list of the tools your team uses, including AI features inside your applicant tracking system. Note what each feature does, which candidates it affects, and where those candidates or roles are located. Then review the relevant vendor documentation, candidate notices, and points where a person checks the result.
