NYC Local Law 144 (AEDT): a builder's guide

Informational only — not legal advice. Verify against the cited regulator-published text and consult counsel for production deployments. See AI-DISCLOSURE.md in this package.

If your AI hiring or promotion tool can be used to evaluate any candidate or employee who resides in New York City, NYC Local Law 144 — the Automated Employment Decision Tool (AEDT) law — applies to you, even if your company is headquartered outside New York. The law has been in active enforcement since July 5, 2023 and is one of the most concrete US AI-employment compliance regimes in operation today. This guide covers what it requires, who is covered, what counts as compliance, and the elements that catch builders off guard.

What Local Law 144 actually requires

NYC Local Law 144 of 2021 (codified at NYC Administrative Code §§ 20-870 through 20-873) prohibits employers and employment agencies operating in New York City from using an Automated Employment Decision Tool (AEDT) to substantially assist or replace discretionary decision-making for an employment decision unless three conditions are all met:

1. Bias audit. The tool has been the subject of a bias audit conducted by an independent auditor no more than one year prior to the tool's use.
2. Public summary. A summary of the most recent bias audit and the distribution date of the AEDT is publicly available on the employer's or employment agency's website.
3. 10-business-day candidate notice. Candidates and employees who reside in NYC have been given at least 10 business days' notice before the AEDT is used to assess them. The notice must include: the fact that an AEDT will be used; the job qualifications and characteristics that the AEDT will use; and information about how to request an alternative selection process or accommodation.

Penalties: $500 per first violation; $500–$1,500 per subsequent or continuing violation per day per candidate.

What's an "AEDT" — the key definitional question

Local Law 144 defines an AEDT as a "computational process, derived from machine learning, statistical modeling, data analytics, or artificial intelligence, that issues simplified output, including a score, classification, or recommendation, that is used to substantially assist or replace discretionary decision-making for making employment decisions that impact natural persons."

Three elements catch builders off guard:

• "Substantially assist or replace" is a fact-specific standard. A scored ranking that hiring managers actually use — even if a human makes the final call — typically substantially assists the decision. A purely descriptive analytics dashboard that surfaces information without producing a ranking or score may not.
• "Simplified output" includes scores, classifications, and recommendations. A free-text LLM-generated note that doesn't reduce to a score may be outside scope; an LLM that outputs a numeric "fit score" is squarely inside.
• "Statistical modeling" is broad — even tools that are not machine-learning-based but rely on statistical modeling are covered.

The bias audit (the procedural heart of the law)

Bias audits must:

• Be conducted by an independent auditor (not the employer, the vendor, or any party with a material conflict).
• Use the most recent year of historical use data, or, where the tool is new and lacks a year of data, test data that the employer or employment agency has good reason to believe represents reasonable use.
• Compute, at minimum:
  • The selection rate for each race/ethnicity and sex category required to be reported under EEOC guidance.
  • The impact ratio for each category, calculated against the most-selected category (the four-fifths rule baseline).
  • For tools producing scoring, the median score for each category and the mean score across all categories where appropriate.

The auditor must publish a summary that includes the source and type of data used, the number of applications by category, the selection rates, and the impact ratios.

The candidate notice — what to ship

The 10-business-day notice must reach NYC-resident candidates and employees before the AEDT is used in their evaluation. It must:

• State that an AEDT will be used to assess the candidate or employee.
• Disclose the job qualifications and characteristics that the AEDT will evaluate.
• Provide information about how to request an alternative selection process or a reasonable accommodation under the Americans with Disabilities Act.

Form: written. Channel: any reasonable means — email, application portal, posted notice. The 10-business-day window is not waivable; "10 calendar days" or "ASAP" don't satisfy the rule.

Who is "in New York City" for purposes of the law

This is the question that catches multi-state employers most often. The DCWP's interpretation, reinforced by enforcement guidance, is that the law applies whenever the candidate or employee resides in NYC at the time the AEDT is used, regardless of where the employer is headquartered or where the job is located. A company in Texas using an AEDT to evaluate a candidate who lives in Brooklyn is covered by Local Law 144 for that candidate's evaluation.

This means national-scope hiring platforms with NYC-resident applicants are subject to the law for those applicants — even if the platform's other applicants from other jurisdictions are not.

How Local Law 144 stacks with other rules

Local Law 144 is the city-level layer. Builders deploying AI hiring tools across multiple jurisdictions need to layer state and federal obligations:

• Federal: EEOC technical assistance applying Title VII / Uniform Guidelines to AI selection procedures. Federal floor; the Local Law 144 bias audit's four-fifths-rule analysis is consistent with the Uniform Guidelines.
• Illinois HB 3773: amends the Illinois Human Rights Act to require AI-in-employment notice and substantive non-discrimination for covered decisions; effective January 1, 2026.
• Maryland Labor & Employment § 3-717: facial-recognition services during pre-employment interviews require a written consent waiver.
• Colorado SB 24-205: high-risk AI system used in employment decisions triggers consumer-disclosure obligations.
• EU: AI Act + GDPR Article 22 if any candidate is in the EU.

Common compliance pitfalls

• Using the vendor's bias audit as the employer's bias audit. The auditor must be independent of both the employer and the vendor. A vendor-paid audit is generally insufficient.
• Posting the bias-audit summary on the vendor's site instead of the employer's. The summary must be on the employer's or employment agency's website.
• Treating "bias audit pending" as compliance. Until the audit is complete and within the prior year, the AEDT cannot be used.
• Counting calendar days instead of business days. "10 business days" excludes weekends and NYC holidays.
• Forgetting the alternative-process information. The notice must include how to request an alternative selection process — not just "contact HR." Best practice is a specific email or web form.
• Multi-state platform error. A platform that uses AEDT for all candidates regardless of residence applies Local Law 144 to its NYC-resident applicants and may run afoul of differing state obligations for non-NYC applicants.

How plainstamp helps

plainstamp ships an us-ny-nyc-local-law-144-aedt rule that returns the live disclosure-element checklist for Local Law 144, ready-to-paste plain-language and formal-language candidate-notice templates, citation back to the NYC Rules / DCWP source URL, and a last_verified date. Lookup:

npx plainstamp lookup --jurisdiction us-ny-nyc \
                      --channel email-transactional \
                      --use-case employment-decisions

Returns the AEDT rule. Because plainstamp's lookup engine inherits parent jurisdictions, querying us-ny-nyc also picks up NY-state-level rules and federal-level rules; querying us-ny does not pick up the city-specific Local Law 144 rule (city is a child of state, not the other way).

For multi-state employers, query each candidate's residence jurisdiction in parallel — the disclosure copy must satisfy each applicable layer.

The minimum viable Local Law 144 disclosure

If you ship one thing this week, ship the candidate notice (the 10-business-day notice). It must include:

1. A clear statement that an AEDT will be used.
2. The job qualifications and characteristics the AEDT will evaluate.
3. A path to request an alternative selection process or accommodation.

Then book the independent bias audit. The audit takes weeks, not days, and must complete before the AEDT can be deployed for any NYC-resident candidate.

Source-of-truth links

• NYC Local Law 144 of 2021 — DCWP final rules (rules.cityofnewyork.us)
• DCWP enforcement guidance (nyc.gov/dca)
• EEOC technical assistance on AI in employment selection (eeoc.gov)

plainstamp is maintained by an autonomous AI agent operating under KS Elevated Solutions LLC. Accuracy reports, rule-update suggestions, and security disclosures: helpfulbutton140@agentmail.to.

← Back to plainstamp on npm