NYC Local Law 144: Bias Audits for AI Hiring Tools
If you use AI anywhere in hiring for New York City roles, Local Law 144 requires an annual independent bias audit and public disclosure. It is already being enforced.
New York City Local Law 144 has been in force since 2023, which makes it different from most of the laws on our regulations map: it is not upcoming, it is actively enforced.
It governs automated employment decision tools, and it applies to a situation many SMEs are already in without realising — using an AI-assisted hiring product for a New York City role.
What the law requires
Two obligations, both concrete:
- An annual bias audit of the tool, conducted by an independent third party.
- Public disclosure of the audit results, plus notice to candidates that an automated tool is being used.
The audit is not a one-off. It has to be repeated every year, and the results must remain publicly available.
Who it catches
The trigger is the role, not the company's headquarters. If you are recruiting for a position based in New York City, the law can reach you from anywhere.
It applies whether you are the tool's deployer (integrating it into your hiring process) or simply using it. In both cases the practical duty is the same: make sure a current audit exists.
The most common way small businesses fall foul of Local Law 144 is not malice. It is buying an applicant-tracking system with AI screening switched on by default, and never asking whether it had been audited.
What to do about it
- Inventory your hiring stack. Applicant tracking, CV screening, ranking, video assessment, chatbot pre-screens — any of these may qualify as automated employment decision tools.
- Ask each vendor for their current bias audit. Reputable vendors in this space publish them. If a vendor cannot produce one, that is your answer about whether to keep using it for NYC roles.
- Check the date. An audit from two years ago does not satisfy an annual requirement.
- Publish the required disclosures and notify candidates before they are assessed.
- Keep records. Being able to show which tool was used, when it was audited, and what candidates were told is the difference between a defensible position and an argument.
What counts as an automated employment decision tool
The boundary is wider than most people assume. It is not limited to systems that reject candidates outright. Tools that substantially assist a decision can qualify — which pulls in:
- CV parsing that scores or ranks applicants.
- Assessment platforms that grade video or written responses.
- Chatbots that screen candidates before a human sees them.
- Matching features inside an applicant tracking system, often enabled by default.
That last one is the trap. Many SMEs never chose an AI hiring tool; they bought recruitment software that added AI features later.
If you cannot say with confidence whether your ATS ranks candidates algorithmically, that is the question to ask your vendor this week.
Why this one matters disproportionately
Hiring is the single most heavily regulated AI use case worldwide. It sits in the high-risk tier of the EU AI Act, it is a consequential decision under the Colorado AI Act, and it triggers discrimination law in nearly every jurisdiction.
If your business uses AI anywhere in recruitment, that is almost always the first use case to bring under proper governance — regardless of which country you operate in.
Frequently asked questions
Does this apply if my company is not based in New York City?
Location of the employer is not the test — the role is. If you are hiring for a position based in NYC, or a remote role tied to NYC, Local Law 144 can apply even if your business sits elsewhere.
Can I rely on my vendor's bias audit?
Often yes, but you remain the party responsible for ensuring an audit exists and is current. Ask your vendor for their most recent audit and confirm it is less than a year old before relying on it.