Automated Employment Decision Tools: Frequently Asked Questions
Citation. NYC Department of Consumer and Worker Protection. (2023, June 29). Automated Employment Decision Tools: Frequently Asked Questions. https://www.nyc.gov/assets/dca/downloads/pdf/about/DCWP-AEDT-FAQ.pdf
Summary. This FAQ explains New York City’s Local Law 144 requirements for automated employment decision tools used in hiring and promotion. It defines covered AEDTs as tools using machine learning, statistical modeling, data analytics, or AI that substantially assist or replace discretionary employment decision-making. The guidance is important for the wiki because it translates algorithmic assessment from a validity/evidence question into a governance question: bias audits, notices, independent auditors, data requirements, and complaints.
Key claims
- Local Law 144 prohibits covered employers and employment agencies from using an AEDT in New York City unless they ensure a bias audit was done and provide required notices.
- The law was enacted in 2021, took effect January 1, 2023, and enforcement begins July 5, 2023.
- An AEDT is described as a computer-based tool that uses machine learning, statistical modeling, data analytics, or AI; helps make employment decisions; and substantially assists or replaces discretionary decision-making.
- Covered employment decisions include screening, not only final hiring or promotion decisions.
- A bias audit is an impartial evaluation by an independent auditor and must include, at minimum, calculations of selection or scoring rates and impact ratios across sex categories, race/ethnicity categories, and intersectional categories.
- The FAQ says the law requires a bias audit but does not itself require specific actions based on bias-audit results; employers and employment agencies remain subject to federal, state, and local anti-discrimination laws.
- Notice requirements apply to NYC resident candidates or employees when a covered AEDT is used.
Evidence & limitations
- This is policy guidance, not empirical evidence about whether bias audits improve fairness, validity, or applicant outcomes.
- The FAQ is jurisdiction-specific to New York City and should not be generalized as a universal legal standard for algorithmic hiring.
- The source is useful for defining governance obligations and compliance vocabulary, especially AEDT, bias audit, independent auditor, impact ratio, screening, and notice.
- The guidance makes adverse impact auditable but does not resolve construct validity, job relatedness, predictive validity, or the adequacy of available audit data.