NYC Local Law 144, Explained: Five Questions Every Employer Using AI in Hiring Asks 

If your company uses any automated tool to evaluate candidates and a New York City resident applies, this law almost certainly applies to you. And as of 2026, the era of light enforcement is over. This blog answers the five questions employers ask. 

1. What is NYC Local Law 144? 

Local Law 144 is the first enforceable AI hiring regulation in the United States. Since July 5, 2023, Local Law 144 has prohibited employers and employment agencies from using an Automated Employment Decision Tool (any AI, machine learning, or statistical process that substantially assists or replaces hiring decisions) unless they complete three requirements: conduct an independent bias audit within the previous 12 months, publish a summary of the results on the company website, and notify candidates at least 10 business days before using the tool. 

The reach is wider than the name suggests. The law follows the candidate, not the employer. A company headquartered anywhere, hiring for a remote role, must comply if the applicant lives in any of the five boroughs. 

2. Why does 2026 change everything? 

Because enforcement just got teeth. In December 2025, the New York State Comptroller released an audit of how the city had been enforcing the law, and the findings were blunt. 75% of complaint calls were misrouted and never reached the enforcement agency. The agency reviewed 32 companies and found one violation; the Comptroller’s auditors reviewed the same companies and found at least 17. 

The agency has committed to proactive enforcement in response, and firms including DLA Piper are warning employers to expect a stricter phase in 2026, with more investigations and higher cumulative penalties. The quiet years created a false sense of security. That window is closing. 

3. What does non-compliance actually cost? 

Penalties start at $500 per violation and escalate to $1,500 per day for ongoing non-compliance, and each use of a non-compliant tool can count as a separate violation. Operating without a valid bias audit for a single month can generate $15,000 to $45,000 in exposure for one tool. For high-volume employers, penalties compound fast. 

• Pro tip: Audits expire after 12 months, and many companies completed their first audit in 2023 and never renewed. If you can’t name your audit’s expiration date right now, that’s your compliance gap. 

4. What counts as an AEDT and what doesn’t? 

Many employers misunderstand what qualifies as an AEDT. An AEDT is any computational process that substantially assists or replaces discretionary hiring decisions by relying exclusively on its output, weighing it above other criteria, or using it to override human judgment. Resume screeners, automated video analysis, and candidate-ranking features in an ATS all qualify. Many companies discover AEDTs inside platforms they assumed were manual, and the obligation falls on the employer, not the vendor. 

The critical distinction is this. Tools where AI makes or dominates the decision are squarely in scope. Tools that document and support a human decision-maker, where the recruiter conducts the interview, exercises judgment, and makes the call, operate on different footing. The law targets the replacement of human judgment, not its support. 

• Pro tip: One question cuts through any vendor’s compliance claims: “who makes the decision, your algorithm or my recruiter?” The answer determines your regulatory exposure more than any feature list.

5. What does defensible look like? 

The employers best positioned for 2026 share a posture, not just paperwork: an inventory of every automated tool in the pipeline, current independent audits published where candidates can find them, documented notice procedures, and, most fundamentally, a hiring process where a human demonstrably leads every consequential decision. 

That last part is the real hedge. Regulation across jurisdictions, including Colorado’s AI Act and the EU AI Act, is converging on the same principle: human oversight of consequential employment decisions. Relevana was built on that principle before it was required. Every interview is recruiter-led, live, and human-decided, independently audited by Asenion across 5,000 candidates with zero disparate impact across gender, race, ethnicity, and every intersectional combination. 

Compliance by architecture, not retrofit. See it live in 20 minutes. Book a Live Demo today.