NYC Local Law 144 — bias audits for AI in hiring.
How Hael produces the statutory bias audit reports and candidate notices NYC Local Law 144 demands — independent audit handoff wired, public summary published.
What NYC LL 144 requires
NYC Local Law 144 regulates Automated Employment Decision Tools (AEDTs) used by employers to screen candidates for jobs in New York City. In force since July 2023, it imposes three substantive obligations: an annual independent bias audit, a candidate-facing disclosure before the AEDT is used, and public publication of a summary of the bias audit results on the employer's or employment agency's publicly accessible website.
The bias audit must be conducted by an independent auditor — a third party with no involvement in the development, deployment or use of the AEDT. The audit must compute selection rates and impact ratios for race, ethnicity and sex categories, following the methodology specified in the DCWP rules. The candidate notice must be provided at least ten business days before the AEDT is used and must disclose the use of the AEDT, the job qualifications and characteristics it assesses, and the source of the data. The Department of Consumer and Worker Protection (DCWP) enforces with civil penalties up to $1,500 per violation per day.
How Hael runs it
Hael wires the independent auditor handoff. Hiring AEDT activity is captured in the agent registry with selection-rate and impact-ratio telemetry continuously collected. When the annual audit cycle opens, scoped read-only access is provisioned for the independent auditor, who can verify the underlying data integrity via hash-chained provenance. The audit summary is generated as a portable document ready for publication on the employer's website.
Candidate notices are generated as portable templates with deployment-context variants (web disclosure, careers-portal banner, written notice). The 10-business-day disclosure window is tracked per requisition. Consent records are sealed in the audit chain. Publication of the public summary is timed against the annual audit cycle and surfaced on the platform's publication calendar.
Questions
What counts as an AEDT under LL 144?
Any computational process derived from machine learning, statistical modelling, data analytics or AI that issues a simplified output (score, classification, recommendation) used to substantially assist or replace discretionary employment decisions. Resume screeners, video interview analysers and predictive hiring tools all qualify.
What independence requirements apply to the auditor?
The auditor must have no involvement in the development, deployment or use of the AEDT; no employment relationship with the employer in the past two years; and no financial interest dependent on the audit outcome. DCWP rules specify the disclosure of independence requirements.
What does the public summary need to include?
Selection rates and impact ratios per race, ethnicity and sex category as computed in the bias audit; the date of the most recent audit; the distribution date of the AEDT version audited; and the source and explanation of the data used. The summary must remain publicly accessible for at least six months.
See Hael run your LL 144 obligations.
Bias audit infrastructure, candidate notice generation, public summary publication — wired against your hiring AEDT inventory.