New York's Local Law 144 vs the EU AI Act: hiring on both sides

Edition 08. A reading time of about 5 minutes.

The hook

If your company hires in both New York City and the European Union, you are under two AI-in-hiring regimes at once, and they are not the same shape. One is narrow, specific, and already being enforced. The other is broad, heavier, and arriving later. The costly mistake is treating either as a substitute for the other. This edition lines them up so you can run one program that satisfies both.

What this means

New York City's Local Law 144 is the narrow, live one. Since enforcement began on 5 July 2023, an employer or employment agency may not use an automated employment decision tool on candidates or employees unless three things are true: the tool has had a bias audit by an independent auditor within the past year, a summary of that audit's results is posted publicly, and the people being evaluated were notified at least ten business days before the tool was used (NYC DCWP, Local Law 144). The enforcement agency can levy penalties from 500 to 1,500 dollars per violation, per day. It is specific, and it is happening now.

The EU AI Act is the broad, later one. It places AI used in recruitment and selection in its high-risk tier (Annex III, point 4), and attaches the full deployer program these editions have walked through: real human oversight, monitoring, record-keeping, and vendor documentation (AI Act, Article 26). Those obligations apply from 2 August 2026 under the Act as published, with a provisional move to 2 December 2027 under an agreement not yet final in the Official Journal, and the fine ceiling runs far higher, up to 15 million euro or 3 percent of worldwide annual turnover (AI Act, Article 99). Broader scope, heavier machinery, later clock.

The trap is assuming one covers the other. A New York bias audit is not an AI Act compliance program; it is one slice of fairness testing, not the oversight, documentation, and monitoring the Act demands. And an AI Act program does not automatically produce the public audit summary or the ten-day candidate notice that New York requires today. They overlap in spirit and diverge in the details that get you fined.

What to do with this

Run one spine, satisfy both heads.

1. For New York, act now. If you use a hiring tool on NYC candidates, get the independent bias audit, post the summary, and send the ten-business-day notice. This is live law with daily penalties, not a 2027 project.
2. For the EU, build the deployer program. Inventory, oversight, monitoring, evidence, the work from earlier editions, on the 2027 clock but worth starting now.
3. Share the evidence between them. The bias audit feeds your AI Act risk story; the candidate notices feed both regimes' transparency expectations. Keep one evidence trail that serves both, rather than two binders that never speak to each other.
4. Track each calendar separately. New York's duty recurs, the audit must stay within the last year, while the EU's obligations arrive on their own date. Same tools, two timelines.

The through-line: regulators on both continents have decided that AI judging people in hiring needs auditing and disclosure. The specifics differ; the direction does not. Build for the direction, and the specifics become a checklist instead of a scramble.

This is educational material and a starting point, not legal advice. It touches two distinct legal regimes, each with its own rules and changes over time, and the EU high-risk timeline rests on an agreement not yet final. For how New York's Local Law 144 and the EU AI Act apply to your specific tools and locations, consult a qualified lawyer in each jurisdiction. Signato is not a law firm and does not certify compliance.

AI that drafts at scale can also send the wrong thing at scale. See the Signato chamber block a forbidden send before it happens: run the free local trial, no card, nothing leaves your machine.