NYC Local Law 144 Requires AI Hiring Audits

New York City’s Local Law 144 requires an independent bias audit, a public summary, and advance notice before an employer or employment agency uses a covered automated employment decision tool (AEDT). Enforcement began July 5, 2023. Coverage depends on the tool’s function and how its output is weighted—not simply whether a product is marketed as “AI.”

What the law covers

The law regulates automated employment decision tools that use machine learning, statistical modeling, data analytics, or AI to issue a score, classification, or recommendation that substantially assists or replaces discretionary hiring or promotion decisions. Under the city’s rule, that means the output is the sole criterion, is weighted more heavily than any other criterion, or is used to overrule or modify a human conclusion. The NYC Department of Consumer and Worker Protection explains the definition and current requirements.

A tool does not become an AEDT merely because it stores applications, searches records, or automates a step. A scoring or ranking model can be covered, but its actual role in the workflow still matters.

Who it applies to

The rule applies to employers and employment agencies using a covered AEDT for candidates for employment or employees for promotion in New York City. Location and remote-work questions can be fact-specific; confirm the current city guidance for the position and people involved.

What employers must do

Three obligations, all before you use the tool:

1. Get an independent bias audit — conducted by an independent auditor no more than one year before use. The required calculations depend on whether the tool selects candidates or assigns scores. See our guide on what a bias audit can and cannot establish.

2. Publish a summary of the audit results on the employer’s or agency’s website. The summary includes the audit date, data source and explanation, sample size, and the required selection or scoring rates and impact ratios.

3. Notify candidates or employees at least 10 business days before use. The notice identifies the AEDT use and the job qualifications and characteristics it assesses. The employer or agency must also publish instructions for requesting an alternative process or accommodation, although Local Law 144 itself does not require one to be granted. Data-source and retention information follows a separate disclosure process under the rule.

What it doesn’t do

It’s worth being precise about the limits, because the law is narrower than the headlines suggested:

• It doesn’t ban biased tools. It requires you to audit, publish, and notify. A tool with a poor impact ratio isn’t automatically illegal under 144 — though it may expose you under other anti-discrimination laws (Title VII, the NYC Human Rights Law).
• The penalties are modest per violation — civil penalties of up to $500 for a first violation and $500–$1,500 for each subsequent one, with each day of use counting separately. The reputational and discrimination-suit exposure usually dwarfs the fine.
• The audit is a snapshot. It reflects the data it was run on. Your live applicant pool can produce different impact, which is why re-auditing on your own deployment matters.

The bigger picture

Local Law 144 is one layer, not the whole map. Illinois regulates certain AI video interviews; other state rules use different definitions and dates; the EU AI Act has its own employment-system framework. Existing federal anti-discrimination law can apply whether or not Local Law 144 covers the tool.

The practical takeaway: document what the system does and how people use its output, then test that workflow against each applicable definition. A product label or the presence of a human reviewer does not answer the coverage question by itself.