NYC Local Law 144, explained for recruiters

What the law requires

New York City enacted Local Law 144 in 2021. It took effect on 1 January 2023, and the Department of Consumer and Worker Protection (DCWP) has enforced it since 5 July 2023. The core rule is short. An employer or employment agency may not use an AEDT to screen a candidate or employee for an employment decision unless:

1. The tool has had a bias audit no more than one year before you use it.
2. A summary of the most recent audit, with the date you started using the tool, is public on your website.
3. Candidates and employees who live in the city got the required notices.

An "employment decision" means screening candidates for employment or employees for promotion. It is not limited to the final offer. DCWP's FAQ says the law applies when a tool substantially helps you assess or screen candidates "at any point in the hiring or promotion process".

What counts as an AEDT

The law defines an AEDT as a computational process "derived from machine learning, statistical modeling, data analytics, or artificial intelligence" that issues a simplified output, "including a score, classification, or recommendation", used to substantially assist or replace discretionary decision making. It excludes tools such as junk email filters, firewalls, antivirus software, calculators, spreadsheets and databases.

DCWP's rules narrow "substantially assist or replace" to three situations, quoted by Gibson Dunn. The FAQ adds a fourth line about sourcing.

On the law's wording, the test turns on that simplified output. A tool that only records and transcribes does not issue a score, classification or recommendation. A tool that scores, ranks or recommends does, and then the question is how much weight your team gives it. Write the answer down for each tool, and review it with counsel.

Who is covered

The law applies to tools used "in the city". DCWP's FAQ says that means any of the following:

• The job is located in an office in New York City, at least part time.
• The job is fully remote, but the office associated with it is in New York City.
• The employment agency using the tool is in New York City, or one of the two points above is true.

If the law applies, the audit must be done before use, and candidates who are New York City residents must get notice. A "candidate" is someone who has applied for a specific position. Employees count when they are being considered for promotion.

The bias audit

A bias audit is "an impartial evaluation by an independent auditor". At a minimum it must calculate selection or scoring rates and the impact ratio across sex categories, race/ethnicity categories and intersectional categories. Under the final rule, as Covington summarizes it, an impact ratio is a category's selection rate divided by that of the most selected category, or its scoring rate divided by that of the highest-scoring category. A ratio below 1 means a group fares worse than the group that fares best.

• Who can audit. DCWP keeps no approved list. An auditor is not independent if they work for you or the vendor, helped use, develop or distribute the tool, or have a financial interest in you or the vendor.
• Which data. Historical data from your own use of the tool. You can rely on an audit pooled across employers only if you gave the auditor your historical data, or it is your first use of the tool. Test data is allowed when historical data is insufficient, and the summary must say why.
• No guessing demographics. Imputed or inferred demographic data cannot be used.
• Small groups. A category below 2% of the audit data can be left out of the calculations.
• How often. Every year. An audit is good for one year from the date it was done.

You must publish the summary on the employment section of your website, or link to it. It must show the audit date, the source and explanation of the data, the number of people in an unknown category, the number of applicants or candidates, the selection or scoring rates, the impact ratios for all categories, and the date you began using the tool.

One point surprises people. The law requires the audit, not any particular response to its results. DCWP's FAQ adds that federal, state and city anti-discrimination laws still apply, so a poor impact ratio is something to act on, not just publish.

The notice: 10 business days before use

Every New York City resident you assess with an AEDT must be told, at least 10 business days before use:

• that an AEDT will be used in assessing them;
• the job qualifications and characteristics the tool will use;
• how to request an alternative selection process or accommodation. DCWP's FAQ says the notice must include instructions for requesting a reasonable accommodation under other laws.

If you do not publish it on your website, you must also provide, within 30 days of a written request, the type of data collected for the tool, its source and your data retention policy.

• Job applicants: notify in the job posting, by mail or email, or on the employment section of your website. A website notice need not be position-specific, and you can start using the tool 10 business days after posting it.
• Promotion candidates: notify by the same channels, or in a written policy or procedure.

The law says the notice must "allow a candidate to request an alternative selection process or accommodation". It does not spell out what the alternative must be. Decide in advance what you will offer, such as an interview with a person, so a request does not stall the process.

Penalties and complaints

Under the law, a violation costs up to $500 for a first violation and each additional violation on the same day, then $500 to $1,500 for each subsequent violation. Each day a tool is used in violation is a separate violation, and so is each missed notice.

Anyone can complain through 311 or DCWP's website. DCWP enforces the audit and notice rules and refers discrimination claims to the NYC Commission on Human Rights. The law also says it does not limit a candidate's right to bring a civil action.

The December 2025 audit: enforcement called "ineffective"

In December 2025 the New York State Comptroller published an audit of DCWP's enforcement from July 2023 to June 2025. As summarized by DLA Piper, it found enforcement "ineffective". DCWP reviewed 32 posted bias audits and found one compliance issue. The Comptroller's team found at least 17 potential issues in the same audits. And 75% of test calls to 311 about AEDTs never reached DCWP.

DLA Piper reads the audit as a signal of more risk for employers, not less. Thin enforcement so far does not change what the law requires, and every day of non-compliant use still counts.

What a vendor can and cannot do for you

DCWP's FAQ is blunt: employers and employment agencies are responsible for making sure a bias audit was done. "The vendor that created the AEDT is not responsible for a bias audit of the tool."

In short, a vendor cannot make you compliant. It can make the work easier, and it should answer three questions in writing: is there a current independent bias audit, whose data did it use, and what exactly does the tool output?

A recruiter's checklist

1. List every tool that touches NYC candidates: sourcing, screening, assessments and interview software.
2. Classify each one. Does it issue a score, classification or recommendation? How does your team use it, against the three situations above?
3. For each AEDT, confirm a bias audit from the past 12 months, by an independent auditor, on data that covers your use.
4. Publish the summary and the date you started using the tool on your careers page.
5. Send notices at least 10 business days before use, and prepare the alternative process you will offer on request.
6. Post your data retention policy, or be ready to send it within 30 days of a written request.
7. Calendar the renewal. The audit lapses after a year.
8. Keep records of the inventory, audits, notices and requests.

If you use Itya, or any interview software, ask the vendor which features produce a score, ranking or recommendation, and decide with counsel how your team will use them. Whether a tool is an AEDT depends on that use, not on the vendor's label. Our trust page lists what we publish about data handling, including our DPA and subprocessors. For the wider picture, see AI interview laws by jurisdiction.