How it works
NYC Local Law 144 mandates that employers using automated employment decision tools (AEDTs) must conduct an annual bias audit to assess whether the tool produces disparate impact across protected groups. The audit must be performed by an independent auditor and published on the employer’s website, with a summary posted in job listings where the tool is used. The law defines disparate impact as a selection rate for any protected class that is less than 80% of the rate for the class with the highest selection rate. The audit must examine outcomes such as hiring, promotion, and termination decisions, and it must include data on the tool’s design, training data, and testing procedures. Employers must retain audit records for three years and provide them upon request to the NYC Department of Consumer and Worker Protection (DCWP).
Also worth reading: How can responsible AI hiring best practices ensure labor law compliance and HR regulatory management in 2026? · Is Your AI Hiring Bias Audit Checklist Ready for 2026 Regulators? · What Are the Best Algorithmic Hiring Audit Standards for Employers in 2026?
The audit process is designed to surface hidden biases that may not be apparent in the tool’s code or documentation. For example, a tool trained on historical hiring data may perpetuate past discrimination if those data reflect biased decisions. The audit must also evaluate whether the tool’s outputs are job-related and consistent with business necessity. While the law does not prescribe specific methodologies, auditors typically use statistical tests such as the four-fifths rule, regression analysis, and fairness metrics like equalized odds or demographic parity. Failure to comply can result in fines of up to $500 per violation per day, and the DCWP has already begun issuing notices to non-compliant employers.
What it costs
NYC Local Law 144 ensures AI fairness in hiring by requiring independent bias audits of automated employment decision tools (AEDTs) before they are used, with results publicly posted. The law mandates that employers and vendors conduct audits to assess whether their AEDTs produce disparate impact based on race, ethnicity, or sex, using statistical thresholds like the four-fifths rule. These audits must be performed by an independent auditor—defined as someone not employed by the employer or vendor—and must evaluate both the tool’s design and its real-world outcomes. By requiring transparency and accountability, the law shifts the burden from job seekers proving discrimination to employers demonstrating fairness proactively. This creates a deterrent effect, encouraging vendors to build more equitable algorithms and employers to scrutinize their tools before deployment. The audit process also includes examining input data, model logic, and output disparities across protected classes, ensuring that bias isn’t just checked at the surface but traced through the system’s mechanics. While enforcement has faced challenges—such as inconsistent audit quality and limited penalties—the law has set a national precedent, pushing other states like Illinois and Connecticut to adopt similar transparency and anti-discrimination measures. Ultimately, Local Law 144 doesn’t just audit code; it audits power, forcing organizations to confront whether their AI tools reinforce or reduce inequity in opportunity. Its strength lies in making fairness a precondition of use, not an afterthought. (178 words)
Common mistakes
Many organizations misunderstand NYC Local Law 144 as a simple checkbox exercise rather than a substantive audit requirement, leading to superficial compliance efforts that fail to address underlying algorithmic bias. A frequent error is relying solely on vendor-provided documentation or internal self-assessments without engaging an independent third-party auditor, which violates the law’s mandate for impartial evaluation. Another common pitfall is conducting the audit too late in the AI lifecycle—after deployment—rather than during design or pre-deployment phases, missing opportunities to remediate discriminatory patterns before they impact hiring outcomes. Some employers also mistakenly believe that using “neutral” data or removing protected characteristics like race or gender from inputs automatically ensures fairness, ignoring proxy variables and systemic biases embedded in historical data. Additionally, failing to publish the audit results conspicuously on the company’s career website for at least six months, or omitting required metrics such as selection rates and impact ratios across demographic groups, undermines transparency and invites enforcement action. These missteps not only risk penalties but also erode trust in AI-driven hiring processes.
To ensure AI fairness under NYC Local Law 144, employers must conduct a rigorous bias audit that evaluates whether their automated employment decision tools (AEDTs) produce disparate impacts based on race, ethnicity, or sex. This involves collecting anonymized applicant and hiring data, calculating selection rates for each demographic group, and computing impact ratios to determine if any group faces a selection rate less than 80% of the highest-performing group—a threshold derived from the Uniform Guidelines on Employee Selection Procedures. The audit must be performed by an independent auditor with no financial or operational ties to the employer or AEDT vendor, ensuring objectivity. Results must include detailed metrics, methodology, and the date of the audit, then posted publicly on the employer’s website. Crucially, the audit is not a one-time task; it must be repeated annually or whenever significant changes are made to the AEDT. By embedding this process into HR technology governance, companies move beyond compliance to build genuinely equitable hiring systems that align with both legal obligations and ethical AI principles.
When to act
NYC Local Law 144 ensures AI fairness in hiring by requiring independent bias audits of automated employment decision tools before they can be used. These audits must assess whether the tool produces disparate impact based on race, ethnicity, or sex, using statistical analysis of selection rates across demographic groups. Employers must publish a summary of the audit results on their website and retain the full report for inspection. The law shifts responsibility from vendors to employers, making it clear that using an un-audited tool—even if developed by a third party—violates the regulation. This creates accountability at the point of deployment, where hiring decisions actually affect candidates. By mandating transparency and empirical validation, the law moves beyond vague promises of “ethical AI” to enforceable, measurable standards. It also encourages vendors to build fairer tools from the outset, knowing non-compliant products will be rejected by NYC employers. The audit requirement acts as a gatekeeper: no audit, no use. This proactive check prevents discriminatory outcomes before they harm job seekers, aligning technological innovation with civil rights protections in one of the nation’s most influential labor markets. Compliance isn’t just legal—it’s a signal of equitable hiring intent.
What to check first
NYC Local Law 144 requires employers using automated employment decision tools (AEDTs) to conduct an annual independent bias audit and publish a summary of the results. The audit must evaluate whether the tool produces disparate impact based on race, ethnicity, or sex, using statistical analysis of selection rates across demographic groups. Employers must first verify that their AEDT falls under the law’s definition—any computational process that screens candidates or employees for hiring, promotion, or other employment decisions. If so, they must engage an independent auditor unaffiliated with the tool’s developer or user to perform the audit. The auditor analyzes historical data to calculate selection rates for each demographic group and identifies whether any group experiences a selection rate substantially less than the most favored group—typically defined as less than 80% of the highest rate, following the EEOC’s four-fifths rule. The audit must also assess data quality, tool transparency, and whether the vendor provided sufficient information for evaluation. Results, including the date of the audit, tool description, data sources used, and selection rate disparities, must be posted publicly on the employer’s website before using the tool. Failure to comply risks civil penalties and reputational harm, making the audit not just a legal obligation but a critical step in ensuring AI-driven hiring practices are fair, accountable, and aligned with anti-discrimination principles under New York City law. Employers should begin by inventorying all AEDTs in use and confirming audit timelines to avoid violations.
How the options compare
| Aspect | NYC Local Law 144 Requirement | Common Audit Practice |
|---|---|---|
| Scope | Applies to automated employment decision tools (AEDTs) used in NYC hiring/promotion | Focuses on algorithmic fairness metrics across race, gender, and ethnicity |
| Audit Frequency | Annual independent bias audit required | Typically conducted yearly or before deployment; some do quarterly checks |
| Transparency | Results must be posted publicly on employer’s website | Often shared internally; public posting varies by jurisdiction and policy |
| Enforcement | NYC Department of Consumer and Worker Protection (DCWP) enforces via fines | Relies on self-reporting and voluntary compliance; limited external oversight |