If you use software to screen, score or rank candidates for jobs in New York City, Local Law 144 requires an independent bias audit every year. Most employers first ask what it costs. The honest answer is that it depends on a handful of things you can control, and knowing them before you ask for quotes saves money and weeks.
Who the law covers
Local Law 144 applies to employers and employment agencies that use an automated employment decision tool (AEDT) for hiring or promotion decisions about candidates and employees in New York City. The Department of Consumer and Worker Protection (DCWP) has enforced it since July 5, 2023.
An AEDT is software built on machine learning, statistical modeling, data analytics or AI that produces a simplified output, such as a score, a classification or a recommendation, and that substantially assists or replaces a discretionary decision. Under DCWP's rules, that means you rely on the output alone, give it more weight than any other factor, or use it to overrule other conclusions, including a human's. Spam filters, spreadsheets and databases are excluded.
What the audit has to contain
- Selection rates for tools that advance or reject people, or scoring rates for tools that score them: the share of each group scoring above the median of the whole sample.
- Impact ratios: each group's rate divided by the rate of the most selected group.
- Results by sex, by race/ethnicity, and by every intersectional category (for example, Hispanic or Latino women).
- The number of people whose category is unknown. They must be counted, not guessed.
- Groups under 2% of the data can be left out of the ratios, but the summary has to say why.
The audit must be done by an independent auditor: someone who did not help build or distribute the tool, does not work for you or the vendor, and has no direct financial interest in either.
When, what to publish, and what to tell candidates
- Timing. The audit can be no more than one year old when you use the tool. In practice, that means once a year, every year.
- Publication. Before you use the tool, post a summary of the results and the date you started using it on the employment section of your website. Keep it up for at least six months after you stop using the tool.
- Notice. Tell NYC candidates and employees at least 10 business days in advance that a tool will be used, what qualifications and characteristics it assesses, and how to request an alternative process or an accommodation.
Penalties
The law sets civil penalties of up to $500 for a first violation and $500 to $1,500 for each one after it. Each day you use a tool without a valid audit, and each missing notice, counts as a separate violation, so the totals add up quickly.
Enforcement so far has been light, but that is changing. In December 2025, the New York State Comptroller published an audit that found DCWP's enforcement ineffective: complaints were misrouted, and in a review of the same 32 published audits, DCWP found one issue while the Comptroller found at least 17. DCWP agreed to strengthen complaint handling and enforcement. Weak enforcement in the past is not a reason to skip the audit now.
What drives the cost of an audit
Auditors usually quote per engagement, so prices vary widely. These five factors decide where you land:
- How many tools. Each AEDT needs its own audit. A résumé screener and a video-interview scorer are two audits.
- Whether you have usable data. Audits are normally based on your historical data from using the tool. Test data is allowed only when there is not enough historical data, and the audit then has to explain why and how the test data was made. Missing or messy data is the most common reason audits get slow and expensive.
- Demographic data. Many employers collect sex and race/ethnicity through voluntary self-identification. Large shares of "unknown" are allowed, but they weaken the results and need explaining.
- Selection or scoring. Tools that only pass or reject are simpler. Scoring tools need the median and scoring rates, and sometimes both.
- Vendor cooperation. If your vendor already provides the output data per candidate, the work is faster. If you have to extract it from an applicant tracking system, budget time for that.
How to keep it cheap
Get your data in order before the auditor starts the clock. Calculate the numbers yourself first, so the audit becomes a check rather than a discovery. If an impact ratio comes out low, you want to know before it is published on your website, not after.
Tools like Fairspire for Local Law 144 calculate these figures inside your own environment, step by step as the DCWP rules describe. If you'd rather have someone do it for you, that is what a pre-audit review is for. A Fairspire Review of one hiring tool costs $1,950 and calculates the selection rates, scoring rates and impact ratios Local Law 144 asks for, including intersectional groups and unknowns, from aggregate data only. It means you walk into the audit knowing your numbers, with tables your auditor can reproduce.
Is a Fairspire Review a certified audit?
No. A Fairspire Review is an independent statistical review of how a decision system treats different groups, with findings and recommendations. It is not a certification, it does not certify compliance with any law, and it does not replace the independent bias audit that Local Law 144 requires. Local Law 144 doesn't set up a certification for auditors either: DCWP does not approve auditors or keep a list of them. What the law asks for is an auditor who is independent.
It is useful for:
- Smaller companies that use an AI or scoring tool but don't have a compliance team or the budget for a large consulting engagement, and want a clear, fixed-price answer on whether their outcomes look fair.
- Anyone who wants to know where they stand against the rules that apply to them, such as Local Law 144 impact ratios or the four-fifths rule, before a candidate, customer or regulator asks.
- Teams preparing for a formal audit: the review gathers the data, tables and documentation the auditor will ask for, so the audit itself is faster, cheaper and has no surprises.
- Teams outside hiring, such as lending, tenant screening or complaint handling, where no law requires an audit yet but disparate impact claims still apply.
- Vendors of AI tools who want to test their product before their customers or auditors do.