NYC Local Law 144 (Automated Employment Decision Tools)

Summary

NYC Local Law 144 prohibits employers and employment agencies from using an Automated Employment Decision Tool (AEDT) to substantially assist or replace discretionary decision-making for hiring or promotion in NYC unless the tool has been subject to an independent bias audit within the past year, the audit summary is publicly posted, and candidates have received required notices. The Department of Consumer and Worker Protection (DCWP) enforces under N.Y.C. Admin. Code §§ 20-870 to 20-874 — effective January 1, 2023, with enforcement deferred to July 5, 2023.

The substantive bias-audit and notice mechanics live in DCWP's Final Rule at 6 RCNY §§ 5-300 to 5-304 (adopted April 6, 2023; retrieved 2026-05-04), which displaced the original 2022 proposed rule and narrowed AEDT scope to tools that substantially assist decision-making — defined as relying solely on the tool's output, weighting it more than any other factor, or using it to overrule alternative human assessments.

What counts as an AEDT in NYC scope

An AEDT is a "computational process derived from machine learning, statistical modeling, data analytics, or artificial intelligence" that produces a "simplified output" — score, classification, or recommendation — used to substantially assist or replace discretionary decision-making. The law's coverage is narrower than its label suggests: under § 20-870 it reaches only hiring or promotion decisions for positions located in NYC or for candidates working in NYC. Performance-management, internal mobility, scheduling, and termination tools fall outside the law unless they directly determine a hire or promote outcome.

The Final Rule's "substantially assist" trigger is the practical filter most enterprise vendors negotiate around: a candidate-scoring system that ranks applicants and feeds a hiring manager is in scope; a resume-parsing tool that extracts text without producing a score, ranking, or recommendation is not. Tools used purely for administrative tasks (interview scheduling, video conferencing, document collection) are out of scope unless they generate scoring or rank-ordering outputs that a decision-maker relies on.

Bias audit requirements

A bias audit must be performed by an independent auditor — defined at 6 RCNY § 5-300 as a "person or group that is not involved in using or developing an AEDT" responsible for the audit. The auditor calculates selection rates (proportion selected per category) and, for tools producing continuous scores, scoring rates (median score per category) across (a) sex categories, (b) race/ethnicity categories drawn from the EEO-1 reporting framework, and (c) intersectional categories combining race/ethnicity with sex. Each rate is compared to the most-selected category to produce an impact ratio, with the federal four-fifths rule (impact ratio < 0.80) referenced as the disparate-impact threshold.

The Final Rule requires audits to use the employer's own historical data when available. Test data is permitted only when the employer has insufficient historical data or the AEDT has been deployed for less than one calendar year. The auditor may exclude from impact-ratio calculations any category representing less than 2% of the audited dataset, but the audit summary must disclose how many candidates fell within excluded or unknown categories. Audits are valid for one year; § 20-871 requires a fresh audit before continued use after that window.

Candidate notice and alternative selection

Under § 20-871(b) and 6 RCNY § 5-303, employers must give candidates and employees who reside in NYC at least 10 business days advance notice before using an AEDT. The notice must (i) state that an AEDT will be used in the assessment, (ii) identify the job qualifications and characteristics the tool will assess, and (iii) include instructions for requesting an alternative selection process or reasonable accommodation. The Final Rule clarifies that posting on a clearly identified AEDT-notice section of the employer's website satisfies the notice requirement; individualized email is not required where the website notice is reasonably accessible.

Employers must accept and respond to alternative-selection requests, but the law does not mandate granting them — only that the request mechanism exist and be accessible. Discrimination against candidates who request the alternative process is independently actionable under federal Title VII and the New York City and State Human Rights Laws.

Penalties and DCWP enforcement

Under § 20-873, civil penalties are $500 for the first violation and for each violation occurring on the same day as the first violation; $500 to $1,500 for each subsequent violation. Each day an AEDT is used in violation of the law constitutes a separate violation, and each missing or non-compliant candidate notice is also a separate violation. A non-compliant deployment running for 30 days exposes an employer to up to $44,000 in stacked daily penalties ($500 day-one violation + 29 × $1,500 daily subsequent violations); a year of non-compliance exposes up to $547,500. DCWP investigates on consumer complaint and ex officio; complaints route through the agency's intake at nyc.gov/dcwp and have not, as of 2026-05-04, produced a published agency enforcement docket.

What NYC LL 144 compliance actually requires

For employers using AEDTs in NYC hiring or promotion, operational compliance has four working parts: (1) commission an independent bias audit covering selection or scoring rates and impact ratios across EEO-1 race/ethnicity, sex, and intersectional categories before first use and annually thereafter; (2) post the audit summary plus the date the AEDT was first used and the data source on a publicly accessible URL on the employer's website; (3) implement a candidate-notice mechanism delivering at least 10 business days advance notice of AEDT use along with the job characteristics being assessed and an accessible mechanism for requesting an alternative selection process; (4) maintain documentation of vendor representations about model inputs, training data, and update cadence — the Final Rule treats the employer, not the vendor, as the responsible party.

LL 144 does not preempt federal Title VII or New York State Human Rights Law disparate-impact obligations. A tool can pass a four-fifths bias audit and still create unlawful disparate impact under finer-grained federal analysis; a tool that fails the LL 144 audit threshold does not automatically violate Title VII. Compliance teams using AEDTs in NYC should treat LL 144 as the floor — federal Title VII pattern-or-practice exposure, EEOC AI guidance on algorithmic discrimination, and DOJ Civil Rights Division algorithmic enforcement priorities sit on top.