NYC Local Law 144 and the Bias Audit, Explained for Recruiters

NYC Local Law 144 is the New York City rule that governs how employers use automated tools in hiring and promotion decisions. If you screen candidates in NYC with any software that helps rank, score, or filter applicants, this law likely applies to you, and it carries real obligations: an independent bias audit, public disclosure of the results, and advance notice to candidates. This guide explains, in plain language, what the law covers, what a bias audit actually involves, and how to stay on the right side of it.

Last updated July 2026. Enforcement update: in December 2025 the New York State Comptroller published an audit of how the Department of Consumer and Worker Protection had enforced Local Law 144 between July 2023 and June 2025, and the agency committed to a more rigorous enforcement posture. Employers who have been treating the bias audit as a low-risk box to leave unticked should expect that assumption to age badly. Penalties start at $500 for a first violation and run to $1,500 for subsequent ones, and each day of non-compliant use can be counted separately. For how this fits with the rules in other states, see our map of AI hiring laws by state.

What the law covers

Local Law 144 regulates what it calls automated employment decision tools, often shortened to AEDTs. Broadly, an AEDT is software that uses machine learning, statistical modeling, or similar techniques to substantially assist or replace a human decision about hiring or promotion. Resume-ranking tools, scoring systems, and automated screeners can all fall under the definition, which means most resume ranking software used on NYC roles is in scope. The law applies when the tool is used to evaluate candidates for positions in New York City, regardless of where the employer is headquartered. The key word is substantially: tools that meaningfully drive the decision are in scope, even when a human signs off at the end.

The three core obligations

For tools in scope, the law sets out three duties.

• An independent bias audit. The tool must be audited within the year before use by an independent auditor who has no stake in its outcome.
• Public disclosure. A summary of the audit results, including the date of the most recent audit, must be published, typically on the employer's careers site.
• Candidate notice. Candidates must be told that an automated tool will be used, given at least ten business days' notice, and informed about the job qualifications and characteristics the tool assesses.

What a bias audit actually involves

A bias audit is a statistical examination of whether the tool produces different outcomes for different groups. The central measure is the selection rate for each group, broken out by sex and by race or ethnicity, and the comparison of those rates. Auditors commonly look at impact ratios, comparing the selection rate of each group to the most-selected group, drawing on the long-established four-fifths convention from federal fairness practice as a screening signal. A meaningful gap between groups is a flag that the tool, or the criteria it applies, may be producing adverse impact and needs attention, and the fix is usually upstream in the criteria themselves, which is what our checklist on reducing bias in AI resume screening works through. The audit is not a one-time certificate. Because tools and applicant pools change, the audit has to be current.

The audit does not ask whether the software is sophisticated. It asks a simpler, harder question: does it select different groups at meaningfully different rates?

Practical steps to stay compliant

Compliance is mostly about process and documentation. Inventory every tool that touches hiring decisions for NYC roles and determine which meet the AEDT definition. Arrange an independent bias audit for each in-scope tool and keep it current. Publish the required summary and date where candidates can find it. Build the candidate notice into your application flow so the timing and content requirements are met automatically rather than remembered case by case. And keep records, because the value of doing this well is being able to show your work if anyone asks.

How a screening-first tool supports compliance

Prescreener was built with this kind of obligation in mind. It is screening-first AI applicant screening that reads the inbound resume pile, applies your criteria, scores role-fit, and ranks candidates, and it is bias-audited to support EEOC and NYC Local Law 144. Because the same criteria are applied consistently to everyone and every ranking shows its reasons, the outputs are the kind of structured, documented decisions that an audit can examine and that you can explain to candidates and regulators alike.

Two design choices matter most for staying compliant. First, AI use is disclosed to candidates, which supports the notice requirement. Second, Prescreener ranks and flags top matches for recruiter review rather than auto-rejecting anyone, keeping a human firmly in the decision. The software supports your obligations, but it does not relieve you of them: you still need your own independent audit, your own published summary, and your own candidate notices in place.

NYC Local Law 144 is less daunting than it first looks once you treat it as a process. Know which of your tools are in scope, keep an independent bias audit current, publish the summary, notify candidates in time, and choose screening software that is transparent, bias-audited, and human-in-the-loop by design. Done that way, compliance and good hiring practice turn out to be the same thing.