NYC Local Law 144: AI in hiring

Checked against Local Law 144 of 2021 as enacted (NYC Administrative Code §§ 20-⁠870 to 20-⁠874), the Department of Consumer and Worker Protection (DCWP) rules, 6 RCNY §§ 5-⁠300 to 5-⁠304, and DCWP’s FAQ of June 29, 2023. As of October 31, 2026, the city’s rules site lists no AEDT rule after the 2023 rules.

Local Law 144 bars employers and employment agencies from using an automated employment decision tool to screen candidates for employment or employees for promotion in New York City unless the tool had a bias audit no more than a year before use and a summary of the results is public. NYC-resident candidates and employees must also get notice (§ 20-⁠871).

Duty	What it requires	Source
Bias audit	An audit by an independent auditor, no more than one year before use	§ 20-⁠871(a); § 5-⁠301
Published results	The audit date, a summary of its results and the tool’s distribution date, posted before use	§ 5-⁠303
Notice	NYC-resident candidates and employees told at least 10 business days before use	§ 20-⁠871(b); § 5-⁠304
Data disclosure	Your data retention policy and the type and source of data posted; written requests answered within 30 days	§ 20-⁠871(b)(3); § 5-⁠304(d)

The law took effect on January 1, 2023, and DCWP has enforced it since July 5, 2023, when its rules took effect (DCWP FAQ I.1).

Independent bias audits and AI governance

Lexara Advisory LLC provides specialized services supporting organizations that use automated employment decision tools (AEDTs) in New York City.

Independent bias audits. Independent bias audits are available through AUDITLL144, a specialized service operated by Lexara Advisory LLC. Each audit engagement is subject to applicable auditor-independence requirements and conflict-of-interest assessments.

AI-governance consultancy. Oblioria offers specialized AI-governance consultancy to employers and employment agencies using automated decision-making technologies. These consultancy services are distinct from independent bias-audit engagements and are subject to appropriate independence safeguards.

Auditor independence and conflict of interest. AUDITLL144 maintains strict auditor-independence standards in accordance with NYC Local Law 144. We do not conduct independent bias audits of automated employment decision tools where Lexara Advisory LLC, through Oblioria or otherwise, or the assigned auditors have participated in the tool’s use, development, or distribution. Every audit engagement is subject to an independence and conflict-of-interest assessment, including relevant employment and financial relationships.

EU AI Act and GDPR. Oblioria’s compliance assessment platform separately addresses the EU AI Act and GDPR, providing expert-reviewed compliance gap reports and access to separately contracted remediation services.

What an AEDT is

An automated employment decision tool (AEDT) is a computational process, derived from machine learning, statistical modeling, data analytics or AI, that issues simplified output, such as a score, classification or recommendation, used to substantially assist or replace discretionary decision making in screening candidates for employment or employees for promotion in the city (§ 20-⁠870).

Under the rules (§ 5-⁠300), “substantially assist or replace” means relying solely on the output, weighting it more than any other criterion, or using it to overrule other conclusions, including human ones. The techniques must generate a prediction or classification, with a computer at least in part identifying the inputs and their weight. Tools that only translate or transcribe, such as converting a resume from PDF, are out. So are tools that do not support or replace discretionary decisions and do not materially impact people, such as spam filters and spreadsheets (§ 20-⁠870).

Who must comply

The duties fall on employers and employment agencies that use an AEDT in the city (§ 20-⁠871). They are not addressed to the vendor that builds the tool; DCWP says the vendor is not responsible for a bias audit of it (DCWP FAQ V.1). An employer can rely on an audit the vendor arranges, but must make sure a valid audit exists (§ 5-⁠301 example). DCWP reads “in the city” as a job at an NYC office at least part time, a fully remote job tied to an NYC office, or an employment agency located in NYC (DCWP FAQ I.4).

The bias audit

Independent auditor. Not independent if it was involved in using, developing or distributing the tool, or, during the audit, works for or has a direct or material indirect financial interest in any employer or agency that uses or plans to use the tool, or in its vendor (§ 5-⁠300).
Timing. No use, or continued use, more than one year after the most recent audit (§ 5-⁠301(a)).
Calculations. For each sex, race/ethnicity and intersectional category from the federal EEO-1 report: the selection rate and the impact ratio against the most selected category (also per group, if the tool sorts people into groups), or, for scoring tools, the share scoring above the median and its impact ratio. The audit reports how many people fell in an unknown category. The auditor may leave a category under 2% of the data out of the impact ratios; the summary then gives the reason and that category’s number of applicants and selection or scoring rate (§ 5-⁠301(d)).
Data. Historical data from the tool’s use. A pooled audit across employers counts for you only if you contributed your data or never used the tool. Test data only when historical data is not enough; the summary then explains why and how the test data was made (§ 5-⁠302).

The law and rules set no pass or fail threshold and require no action on the results, but federal, state and city anti-discrimination laws still apply (DCWP FAQ II.2).

Publishing results and giving notice

Results. Before use, post clearly and conspicuously on the employment section of your website, or link to, the date of the most recent audit, a summary of its results (data source, unknown count, and per category the number of applicants, rates and impact ratios) and the tool’s distribution date. Keep them up for at least six months after the tool’s latest use (§ 5-⁠303).
Notice. At least 10 business days before use, tell NYC-resident candidates and employees that an AEDT will be used and which qualifications and characteristics it will assess (§ 20-⁠871(b)). Include how to request an alternative process or an accommodation, if available; you need not offer an alternative (§ 5-⁠304(a)). Candidates can be notified on your website, in the job posting or by mail or email; employees up for promotion in a written policy, the job posting or by mail or email (§ 5-⁠304(b)-(c)).
Data disclosure. Post your AEDT data retention policy and the type and source of data on the employment section of your website, with instructions for written requests, and answer a request within 30 days, or explain why disclosure would break the law or interfere with a law enforcement investigation (§ 5-⁠304(d)). The law itself makes this information available on written request where it is not on your website (§ 20-⁠871(b)(3)); the rule requires both.

Penalties

Up to $500 for a first violation and each additional violation that day, then $500 to $1,500 for each later one. Each day of unlawful use is a separate violation, and so is each missed notice (§ 20-⁠872(a)-(c)). Penalty cases go to a tribunal within the city’s Office of Administrative Trials and Hearings (OATH) or within another city agency designated to hear them (§ 20-⁠872(d)), and the Corporation Counsel can sue for compliance (§ 20-⁠873). The law leaves candidates’ and employees’ own lawsuits and the NYC Commission on Human Rights’ powers intact (§ 20-⁠874).

Hiring in the EU too? AI used to recruit or select people, or to decide on promotion or termination, is listed as high-risk under the EU AI Act (Annex III point 4), with duties from December 2, 2027. See high-risk AI system and The EU AI Act for US companies.
A company that uses a hiring tool someone else built is usually its deployer under the EU AI Act: Provider or deployer: your EU AI Act role.
DCWP’s AEDT page has the official guidance and how to file a complaint.

Sources

Local Law 144 of 2021, enacted text (NYC Council)
DCWP rules, 6 RCNY §§ 5-⁠300 to 5-⁠304: Notice of Adoption and rule page (effective date)
DCWP, AEDT frequently asked questions (June 29, 2023)
DCWP, Automated Employment Decision Tools

This page provides general regulatory information. Oblioria does not provide legal advice on New York law.