Loading…
Compliance
New York City Local Law 144 requires employers using Automated Employment Decision Tools (AEDTs) to commission an independent bias audit before use, publish the results publicly, and notify candidates at least 10 business days before the tool is applied to their application. Violations carry fines of $375–$1,500 per day, per violation.
Most teams building AI hiring tools know the law exists but aren't clear on what the technical compliance work actually entails. The audit is not a questionnaire. It requires a statistical analysis of disparate impact ratios across intersectional demographic categories, a documented methodology, and a test dataset with sufficient demographic representation. We build the pipeline, produce the documentation, and prepare you for an independent auditor to sign off.
Tell us what you're building.
Five requirements that any employer using an AEDT for NYC hiring or promotion decisions must satisfy. Each has technical implications beyond the legal obligation.
Employers and employment agencies must commission an independent bias audit before using any Automated Employment Decision Tool (AEDT). The audit must calculate selection rate and score rate by sex, race/ethnicity, and intersectional categories. “Independent” means the auditor has no financial relationship with the employer or the AEDT vendor beyond the audit engagement itself.
The summary of audit results must be publicly available on the employer's website before the AEDT is used in any hiring or promotion decision. The summary must include the date of the audit, the name and contact information of the independent auditor, the categories of data used to train or test the AEDT, the number of individuals in each category, and the selection or score rates for each category.
Employers must notify candidates and employees that an AEDT will be used in the hiring or promotion process at least 10 business days before the tool is used. Notice must be given via the job posting or by personal notice (email or in person). Candidates must also be informed that they may request an alternative selection process or accommodation.
Audit data, methodology documentation, and the results of any bias audit must be retained for at least 3 years. This includes the dataset used for the audit, the scoring distributions by demographic category, and any remediation actions taken in response to audit findings.
A new bias audit is required at least annually, or whenever there is a material change to the AEDT, including changes to the underlying model, training data, or scoring criteria. Employers who continue using an AEDT beyond 12 months without a new audit are in violation of the law.
The specific technical deliverables for LL144 compliance. These cover both the bias audit preparation and the operational requirements that must run annually.
A technical pipeline that calculates selection rates and score rates by sex, race/ethnicity, and intersectional combinations (e.g., Black women vs. white men). The pipeline produces the specific metrics required by the law's audit standard: selection rate ratio, score rate ratio, and statistical significance testing for observed differences across demographic groups.
The audit dataset must include sufficient representation of each demographic category to produce statistically meaningful results. We design the test dataset stratification, identify gaps where demographic categories are underrepresented, and document the methodology so an independent auditor can verify and replicate the analysis.
We produce the technical documentation package that an independent auditor needs to conduct their review: model description, training data summary, scoring criteria documentation, test dataset statistics, and the full results with methodology. The independent auditor conducts the formal review. We give them everything they need to do it efficiently.
Technical implementation of the 10-business-day candidate notice requirement, automated notice via job application email sequences or applicant tracking system (ATS) integrations. Includes opt-out and alternative process request handling, with logging of notice delivery and candidate responses for audit trail purposes.
The bias evaluation pipeline uses open-source fairness toolkits that produce auditable, reproducible results. Candidate notice automation integrates with your existing ATS.
We work with your current ATS rather than requiring you to switch platforms. If your ATS is not listed here, ask us. We have integrated with most major platforms.
ATS with API access for candidate notice automation. Supports custom email triggers at application stage.
ATS with webhook support for notice trigger automation. Data export APIs support audit dataset preparation.
Enterprise ATS with AEDT disclosure workflow capabilities. Common in enterprise LL144 compliance programmes.
Open-source bias and fairness toolkit from the University of Chicago. Used for disparate impact ratio calculation and audit pipeline construction.
Open-source fairness assessment library for machine learning models. Produces the group metric comparisons required for LL144 audit documentation.
LL144 compliance work has a specific scope. We decline engagements that are outside it.
Hiring tools used exclusively outside New York City
NYC Local Law 144 applies when an AEDT is used to evaluate candidates for positions in New York City, or to evaluate employees based in New York City for promotion. If all of your hiring and promotion decisions affect positions located outside NYC, LL144 does not apply. Note that remote positions where the candidate could work in NYC may still be in scope, confirm with your employment counsel.
Tools that are advisory only and do not substantially assist in decisions
LL144 applies to AEDTs that “substantially assist or replace discretionary decision-making” in hiring or promotion. A tool that provides information to a human who makes an independent judgement may not qualify as an AEDT. Whether your tool meets this threshold is a legal question, your employment counsel should make that determination. We build the technical compliance infrastructure once the determination is made.
Teams expecting a compliance opinion from us
We build the bias evaluation pipeline and audit documentation package. Whether your tool is an AEDT under the law, whether your audit satisfies the legal standard, and whether your notice procedures comply. Those are legal questions for your employment counsel. We produce the technical evidence; your legal team determines whether it satisfies the law's requirements.