NYC Local Law 144: The First Mandatory AI Bias Audit Law in the United States

Written by

in

NYC Local Law 144: The First Mandatory AI Bias Audit Law in the United States

Since July 5, 2023, employers in New York City using automated employment decision tools (AEDTs) to screen candidates or employees must commission annual independent bias audits and publish the results — the first U.S. law to mandate algorithmic auditing of hiring tools.

What the Law Requires

NYC Local Law 144, passed in December 2021 and effective July 5, 2023, requires that:

  1. An employer or employment agency may not use an automated employment decision tool (AEDT) to screen a candidate or employee for employment unless the tool has been subject to an annual independent bias audit.
  2. The results of the bias audit must be published on the employer’s website.
  3. Candidates and employees must be notified at least 10 business days before the AEDT is used, including information about the data sources used, the qualifications assessed, and the retention period.

What Counts as an AEDT

An AEDT is defined as any computational process derived from machine learning, statistical modeling, data analytics, or artificial intelligence that:

  • Issues a simplified output (score, tag, classification, recommendation) AND
  • Is designed to substantially assist or replace discretionary decision-making in hiring

Notably, the final rules clarified that a tool “substantially assists” decision-making if it relies on a score or ranking to select candidates, even if a human reviews the results. This addresses the “human-in-the-loop” question — a rubber-stamp review does not exempt the tool from the audit requirement.

The Bias Audit

The independent audit must evaluate the tool’s selection rates across protected categories (sex, race/ethnicity) and calculate impact ratios in compliance with the Equal Employment Opportunity Commission’s Uniform Guidelines on Employee Selection Procedures (29 C.F.R. § 1607). The audit must include:

  • The number and demographic categories of individuals the tool assessed
  • Selection rates for each demographic category
  • Impact ratios comparing selection rates across categories
  • The distribution of candidates across categories

Enforcement

The NYC Department of Consumer and Worker Protection (DCWP) enforces the law. Violations carry civil penalties of:

  • $500 for the first violation
  • $500–$1,500 for each subsequent violation (per day)

The law does not provide a private right of action.

How It Compares

Local Law 144 is narrower than Colorado’s AI Act (SB 24-205), which covers all high-risk AI systems — not just hiring tools — and requires impact assessments and risk management plans. But Local Law 144 is more specific in its audit requirements: it mandates an independent third-party audit, while Colorado’s law allows self-assessment.

The EU AI Act similarly classifies AI systems used in employment as “high-risk” (Annex III), requiring conformity assessments, risk management systems, and human oversight — but does not mandate independent bias audits in the same way.

Primary Sources


AI automated and generated content. Human reviewed. All policy text is summarized from the primary sources linked above.

The Weekly AI Policy Brief

AI policy moves fast. Get the week’s developments, new trackers, and podcast episodes in one short email — free, ad-free, no spam.

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *