NYC Local Law 144: AEDT Guide
If you use AI to hire or promote employees in New York City, you are subject to Local Law 144. Here is how to conduct your bias audit.
What is Local Law 144?
Effective July 5, 2023, NYC Local Law 144 prohibits employers from using an Automated Employment Decision Tool (AEDT) unless:
- The tool has been the subject of a bias audit within the last year.
- The results of the audit are made publicly available.
- Candidates are given notice that an AEDT is being used.
What Counts as an AEDT?
An AEDT is any computational process derived from machine learning, statistical modeling, or data analytics that issues a "simplified output" (score, tag, classification) used to substantially assist or replace discretionary decision making.
Examples: Resume scanners, AI interview bots, candidate ranking algorithms.
The Bias Audit Requirement
The audit must be conducted by an independent auditor. It must calculate the "Selection Rate" and "Impact Ratio" for each category (race, ethnicity, sex).
Impact Ratio Formula
Impact Ratio = (Selection Rate of Group) / (Selection Rate of Most Selected Group)If the Impact Ratio is less than 0.8 (the "four-fifths rule"), it may indicate adverse impact, though the law mandates disclosure rather than a specific passing score.
Railguard for HR Tech
Railguard helps HR technology vendors and employers maintain continuous compliance with NYC 144.
- Continuous Monitoring: We track selection rates in real-time, alerting you if your model starts drifting into bias territory.
- Audit Logs: We maintain a cryptographic record of every decision made by the model, which is essential for the independent audit.
- Explainability: Generate "Why" explanations for every candidate ranking to prove decisions were not based on protected class variables.
Bias Audit Checklist
Download our checklist to ensure your AEDT audit meets all NYC requirements.