NYC Local Law 144, effective July 5, 2023, requires employers using automated employment decision tools to conduct annual independent bias audits, publish results publicly, and provide candidates 10 business days written notice before using AEDT.
The law applies to employers and employment agencies using automated employment decision tools (AEDT) in hiring or promotion decisions affecting New York City candidates or employees. Both the employer and the AEDT vendor can face liability. Violations carry penalties of $375 for a first offense and $1,500 for each subsequent violation, assessed per person per day of non-compliance, enforced by the NYC Department of Consumer and Worker Protection (DCWP).
Who This Applies To: Employers and employment agencies using automated employment decision tools (AEDT) in hiring or promotion decisions affecting NYC candidates or employees. An AEDT is any computational process, derived from machine learning, statistical modeling, data analytics, or artificial intelligence, that issues a simplified output used to substantially assist or replace discretionary decision-making. Both the employer deploying the AEDT and the vendor providing it can face liability for non-compliance.
New York City Local Law 144 was the first law in the United States to regulate the use of artificial intelligence in employment decisions. Enacted on December 11, 2021, and effective July 5, 2023, the law requires employers using automated employment decision tools to conduct annual independent bias audits, publish the audit results on their website, and provide candidates with advance written notice before the tool is used. The law specifically addresses hiring and promotion decisions, creating concrete obligations for any employer using AI screening, scoring, or ranking tools in the talent pipeline.
Local Law 144 has served as a model for AI hiring regulation across the United States, with several states and municipalities considering similar legislation. The law's narrow focus on employment AI makes it operationally specific: employers know exactly what they must do, when they must do it, and what the consequences are for non-compliance. For how this law fits into the broader AI regulatory landscape, see our 2026 AI regulatory compliance guide. For AI governance in HR specifically, see our AI governance for HR guide.
This guide covers the law's requirements in detail, including the AEDT definition, bias audit specifications, candidate notice obligations, the penalty structure, and the compliance steps employers need to take. The law's seemingly straightforward requirements conceal operational complexity, particularly around what constitutes an AEDT, what a compliant bias audit looks like, and how to operationalize the 10-business-day notice requirement at scale.
What NYC Local Law 144 Requires
Automated Employment Decision Tool (AEDT) Definition
An AEDT is defined as any computational process, derived from machine learning, statistical modeling, data analytics, or artificial intelligence, that issues a simplified output including a score, classification, or recommendation, that is used to substantially assist or replace discretionary decision making for making employment decisions that impact natural persons. The key phrase is "substantially assist or replace." If an AI tool provides a score or ranking that a human hiring manager can override but generally follows, the tool likely qualifies as an AEDT. Simple keyword filters or boolean searches that do not use machine learning or statistical modeling generally do not qualify. The distinction matters because the entire compliance obligation applies only to tools meeting the AEDT definition.
Annual Independent Bias Audit
Before using an AEDT and annually thereafter, the employer must ensure the tool has been subject to an independent bias audit conducted no more than one year before the date of use. The bias audit must be conducted by an independent auditor who is not involved in developing or deploying the AEDT. The audit must test the tool for disparate impact on the basis of sex, race/ethnicity, and intersectional categories of sex and race/ethnicity. The audit must calculate the selection rate and impact ratio for each demographic category. The impact ratio is the selection rate for a given category divided by the selection rate for the most selected category. If the AEDT selects candidates for further consideration, the audit analyzes selection rates. If the AEDT scores or classifies candidates, the audit analyzes scoring rates using either the average score per category or the proportion scoring above the median. The audit must use historical data from the employer's own use of the tool. If the employer lacks sufficient historical data, the audit may use test data, but the summary must disclose this limitation.
Public Disclosure of Audit Results
The employer must make the summary of the most recent bias audit results publicly available on the employer's website. The summary must include the source and explanation of the data used in the audit, the number of individuals assessed by the AEDT, the selection rates and impact ratios for each demographic category, and the date of the audit. The summary must remain posted on the website for at least six months after the latest use of the AEDT in an employment decision. This public disclosure requirement creates transparency pressure beyond the audit itself because the results are visible to candidates, competitors, regulators, and the media.
Candidate Notice Requirements
Employers must provide candidates with written notice at least 10 business days before the AEDT is used. The notice must state that an AEDT will be used in connection with the employment decision and must describe the job qualifications and characteristics that the AEDT will use in its assessment. The notice must also inform the candidate of their right to request an alternative selection process or accommodation and must provide instructions for how to make such a request. Separately, candidates must be notified about the data collected and the data retention policy, including information about the type of data collected, the source of the data, and the employer's data retention policy. Candidates who are NYC residents must be informed of their rights under the NYC Fair Chance Act if applicable.
Recordkeeping
While Local Law 144 does not prescribe specific recordkeeping requirements beyond the public audit summary, the DCWP's enforcement approach and the operational needs of compliance create de facto recordkeeping obligations. Employers should maintain records of all bias audits including the full audit report (not just the summary), documentation of candidate notices sent including dates and delivery confirmation, records of any accommodation or alternative process requests and how they were handled, documentation of the AEDT vendor's representations about the tool's functionality, and any changes made to the AEDT or its configuration since the last bias audit.
Key Dates and Enforcement Timeline
| Date | Requirement | Who | Status |
|---|---|---|---|
| December 11, 2021 | Local Law 144 enacted | All NYC employers using AEDT | Complete |
| April 6, 2023 | DCWP final rules published defining AEDT, audit requirements, and notice procedures | All NYC employers using AEDT | Complete |
| July 5, 2023 | Local Law 144 takes effect; all requirements enforceable | Employers and employment agencies | Active |
| July 5, 2023 | Bias audits must have been completed within 1 year before AEDT use | Employers | Active |
| 2024-2025 | DCWP begins enforcement; first complaint investigations and inspections | All covered employers | Active |
| 2026 Ongoing | Annual bias audit renewals due for employers who first audited in 2023; DCWP enforcement escalation | All covered employers | Active |
Penalties for Non-Compliance
Local Law 144 is enforced by the NYC Department of Consumer and Worker Protection (DCWP). The DCWP can initiate investigations based on complaints or on its own initiative and can conduct inspections of employer records. Penalties are structured as civil penalties assessed per violation.
For a first violation, the penalty is $375. For each subsequent violation, the penalty is $1,500. Critically, each use of a non-compliant AEDT on a candidate constitutes a separate violation, and each day of continued non-compliance constitutes an additional violation. This per-person, per-day structure means penalties accumulate rapidly for employers processing large volumes of candidates. An employer screening 100 candidates per day without a valid bias audit could face $37,500 in penalties for the first day alone (100 candidates times $375) and $150,000 per day thereafter (100 candidates times $1,500).
The failure to publish the bias audit summary on the employer's website is a separate violation from the failure to conduct the audit and the failure to provide candidate notice. An employer could face concurrent penalties for all three failures. The DCWP can also issue cease-and-desist orders requiring the employer to stop using the non-compliant AEDT until the employer demonstrates compliance, which disrupts recruiting operations.
Both the employer deploying the AEDT and the employment agency or vendor involved in providing the tool can face liability under the law. This shared liability model creates incentive for employers to verify vendor compliance and for vendors to support their customers' audit and notice obligations. The law also does not preempt other applicable laws, meaning an employer using a biased AI hiring tool could face simultaneous liability under Local Law 144, federal Title VII disparate impact claims, and applicable state anti-discrimination statutes.
Compliance Checklist
- ☐ Audit all hiring and promotion tools to determine which meet the AEDT definition (computational process using ML, AI, or statistical modeling that substantially assists or replaces discretionary decisions)
- ☐ Engage an independent auditor to conduct bias audits for each identified AEDT, testing for disparate impact by sex, race/ethnicity, and intersectional categories
- ☐ Publish the bias audit summary on your company website including data source, sample sizes, selection rates, impact ratios, and audit date
- ☐ Implement a 10-business-day advance written notice process for all candidates who will be assessed by an AEDT
- ☐ Include in candidate notices a description of the job qualifications and characteristics the AEDT evaluates
- ☐ Build a process for candidates to request alternative selection processes or accommodations and track all requests and resolutions
- ☐ Provide separate data collection and retention notices to candidates describing the types and sources of data collected by the AEDT
- ☐ Establish an annual bias audit renewal calendar to ensure audits are refreshed before the one-year expiration
- ☐ Maintain comprehensive records including full audit reports, notice delivery confirmations, accommodation requests, and AEDT vendor documentation
Local Law 144 compliance requires coordination between HR, legal, talent acquisition, and your AEDT vendors. The 10-business-day notice requirement alone requires changes to recruiting workflows and applicant tracking systems. Contact PolicyGuard to see how we can help you operationalize Local Law 144 compliance across your hiring process.
PolicyGuard helps companies like yours get AI governance documentation audit-ready in 48 hours or less.
Start free trial →How PolicyGuard Helps
PolicyGuard provides specific capabilities designed for Local Law 144 and AI hiring law compliance:
- AEDT Identification and Classification: PolicyGuard analyzes your hiring technology stack to identify which tools meet the AEDT definition under Local Law 144. Many organizations use AI-powered features within broader applicant tracking systems without realizing those features qualify as AEDTs. PolicyGuard maps each tool's functionality against the legal definition and flags those requiring bias audits and candidate notices.
- Bias Audit Management: PolicyGuard manages the bias audit lifecycle, from auditor selection through audit completion and annual renewal. The platform tracks audit expiration dates, alerts your team when renewal is approaching, and stores full audit reports and public summaries with version history. When your auditor delivers results, PolicyGuard helps you assess whether impact ratios indicate disparate impact requiring remediation before the results are published.
- Candidate Notice Automation: PolicyGuard integrates with applicant tracking systems to automate the 10-business-day notice requirement. The platform generates compliant notice language based on each AEDT's specific characteristics, tracks delivery and acknowledgment, and ensures no candidate is assessed by an AEDT before the notice period has elapsed. This prevents the most common compliance failure: using the AEDT before the notice period is complete.
- Alternative Process Request Tracking: PolicyGuard tracks candidate requests for alternative selection processes and accommodations, routing requests to the appropriate HR team members, tracking resolution, and documenting outcomes. This creates an audit trail demonstrating that the employer takes accommodation requests seriously and processes them consistently.
- Compliance Evidence Packages: PolicyGuard generates Local Law 144 compliance evidence packages for DCWP inspections, including bias audit documentation, notice delivery records, accommodation request logs, and vendor compliance certifications. When the DCWP requests documentation, your HR and legal teams can produce a complete evidence package within hours rather than weeks.
FAQ
Does Local Law 144 apply to remote positions that NYC residents might apply for?
Local Law 144 applies to employment decisions that impact candidates or employees in New York City. If a remote position could be filled by an NYC resident and the employer uses an AEDT to screen candidates, the law applies to the extent the tool is used on NYC candidates. The DCWP rules clarify that the law covers AEDTs used to evaluate candidates for employment or employees for promotion within New York City. Employers should apply Local Law 144 compliance to any hiring process where NYC-based candidates may be in the applicant pool.
Who can conduct the independent bias audit?
The bias audit must be conducted by an independent auditor who has not been involved in using, developing, or distributing the AEDT. The law does not specify auditor qualifications or certifications, but the auditor must be independent of the employer and the AEDT vendor. In practice, this means third-party firms specializing in algorithmic auditing, industrial-organizational psychology firms, or data analytics consultancies that can conduct statistical disparate impact analyses. The employer should ensure the auditor has demonstrable expertise in bias testing methodology and understands the specific statistical requirements of the DCWP rules.
What if the bias audit reveals disparate impact?
The law does not prohibit the use of an AEDT that shows disparate impact. It requires the audit to be conducted and the results to be published. However, using a tool that you know produces disparate impact creates significant legal risk under federal Title VII, the New York State Human Rights Law, and the NYC Human Rights Law. If the audit reveals impact ratios below 0.8 (the four-fifths rule threshold commonly used in employment discrimination analysis), the employer should work with the AEDT vendor to understand the source of the disparity, explore recalibration or alternative tools, and consult employment counsel about the legal risk of continued use. Publishing results showing significant disparate impact also creates public relations and recruiting brand risk.
How do we handle the 10-business-day notice for high-volume recruiting?
The 10-business-day notice requirement is one of the most operationally challenging aspects of the law. For high-volume recruiting, employers typically integrate the notice into the application process itself. When a candidate applies, the notice is provided as part of the application workflow, and the AEDT is not applied until 10 business days have elapsed. Some employers restructure their recruiting timeline to include an initial human review phase during the notice period, with AEDT screening beginning only after the notice window closes. Applicant tracking system configuration is critical: the system must track when notice was provided and prevent AEDT processing before the 10-day window has passed.
Are AI tools used only for sourcing or initial outreach covered?
The AEDT definition focuses on tools that substantially assist or replace discretionary decision-making in employment decisions. AI tools used solely for sourcing candidates, such as identifying potential candidates on LinkedIn, generally do not meet the AEDT definition because they assist in outreach rather than in making employment decisions. However, if a sourcing tool scores or ranks candidates and that scoring substantially influences which candidates move forward in the hiring process, it may qualify as an AEDT. The line between sourcing and screening is fact-specific. If the tool's output determines whether a candidate receives consideration, it likely qualifies. If the output merely identifies a pool of potential candidates for human review with no ranking or scoring, it likely does not. When in doubt, apply the law's requirements to avoid enforcement risk. See our AI policy governance guide for additional guidance on classifying AI tools in employment contexts.
NYC Local Law 144 set the standard for AI hiring regulation in the United States, and its requirements are becoming the baseline that other jurisdictions build on. Employers who build robust Local Law 144 compliance programs now will have a strong foundation when similar laws emerge in other cities and states. Talk to PolicyGuard about building a hiring AI compliance program that meets NYC requirements and scales nationally.
