How to Challenge an AI Hiring Tool in New York City Under Local Law 144 and Anti-Discrimination Rules
New York City Local Law 144 requires employers using automated employment decision tools (AEDTs) to conduct an annual “bias audit” and publish key audit details before using the tool on NYC applicants. In practice, many employers still deploy AI screening systems that can disadvantage protected groups or violate notice rules. This article explains how to identify violations, preserve evidence, and challenge an AI hiring tool under Local Law 144, the NYC Human Rights Law, and other anti-discrimination laws.
What NYC Local Law 144 Regulates (and What It Doesn’t)
New York City Local Law 144 (the “AI hiring law”) targets the growing use of automated scoring and ranking systems in employment. The law regulates an employer’s or employment agency’s use of an “automated employment decision tool” (AEDT) when the tool is used to substantially assist in making hiring or promotion decisions for NYC candidates.
In plain terms: if software helps decide who advances, who is rejected, or how applicants are ranked—based on algorithmic processing—Local Law 144 may apply.
Key compliance duties under Local Law 144
When covered, the employer (or employment agency) must generally:
(1) Conduct a “bias audit” within the past year before using the AEDT for covered decisions. A bias audit is an independent evaluation intended to assess disparate impact by sex, race/ethnicity, and other categories used in audit reporting.
(2) Publish a summary of the bias audit on a website, including specific output metrics in a publicly accessible format.
(3) Provide required notices to candidates and employees in New York City that an AEDT will be used, identify job qualifications and characteristics the AEDT will assess, and provide information about requesting an alternative selection process or accommodation where applicable.
Important limits: Local Law 144 is not a general “AI fairness” statute
Local Law 144 can be a powerful entry point, but it does not automatically create a discrimination claim. A tool can be “audited” and still produce discriminatory outcomes in practice, and conversely a notice violation can occur without any discriminatory intent. For many clients, the best strategy is to use Local Law 144 to obtain leverage and facts, while pursuing substantive anti-discrimination claims under the NYC Human Rights Law (NYCHRL), New York State Human Rights Law (NYSHRL), and federal law (including Title VII and the ADA).
Spotting Red Flags: Signs an AI Hiring Tool May Be Challengeable
AI-driven screening can create legal exposure when it:
Rejects candidates rapidly after online assessments, recorded interviews, résumé parsing, or chat-based screening—with little explanation.
Uses proxies for protected traits (e.g., zip code, school, employment gaps, speech patterns) that correlate with race, national origin, disability, age, or sex.
Penalizes disability-related traits (e.g., monotone speech, atypical eye contact, processing speed, assistive technology use).
Undervalues nontraditional career paths common among protected groups (e.g., caregiving gaps affecting women; interrupted employment due to disability).
Fails to provide Local Law 144 notices or provides vague, late, or buried notices that do not meaningfully inform the candidate.
Example: Automated video interview scoring
A candidate completes a recorded interview where software analyzes word choice, tone, pace, and facial expressions. The system flags the applicant as “low enthusiasm” and rejects them. If the candidate has a speech disability, autism, or uses assistive technology, the tool may have screened out protected disability traits—supporting an ADA/NYCHRL disability discrimination theory. If the employer did not give compliant Local Law 144 notice and did not publish a valid bias audit summary, those procedural violations can strengthen a challenge and prompt early settlement.
Legal Framework: Claims That Pair With Local Law 144
Local Law 144 often works best when combined with broader anti-discrimination statutes that address disparate treatment, disparate impact, failure to accommodate, and retaliation.
1) NYC Human Rights Law (NYCHRL)
The NYCHRL is among the most protective anti-discrimination laws in the country. It covers employers in NYC and prohibits discrimination in hiring based on protected characteristics including race, national origin, sex, gender identity, disability, age, religion, sexual orientation, arrest/conviction record (with specific rules), and more.
NYCHRL claims can be especially viable where an AEDT “substantially assists” a decision and produces biased outcomes, or where the employer fails to reasonably accommodate a disability during an AI-driven assessment.
2) New York State Human Rights Law (NYSHRL)
NYSHRL also prohibits employment discrimination statewide. Depending on the facts, plaintiffs may plead both NYCHRL and NYSHRL claims, with NYCHRL often providing broader protections and remedies.
3) Federal law: Title VII and ADA
Title VII prohibits discrimination based on race, color, religion, sex, and national origin. AI screening tools are a common factual setting for disparate impact claims—where a neutral practice disproportionately excludes a protected group and is not job-related and consistent with business necessity (or where a less discriminatory alternative exists).
The ADA prohibits disability discrimination and requires reasonable accommodations in hiring processes. AI assessments that penalize disability-related communication or testing conditions can trigger ADA liability, particularly if the applicant requested an accommodation and was denied or ignored.
4) EEOC attention to algorithmic decision-making
The Equal Employment Opportunity Commission has repeatedly emphasized that employers remain responsible for discrimination caused by third-party vendors’ hiring algorithms. That principle matters in litigation: an employer generally cannot outsource compliance by blaming the vendor’s “black box.”
Step-by-Step: How to Challenge an AI Hiring Tool in NYC
Step 1: Determine whether Local Law 144 likely applies
Key factual questions include:
Was the role in New York City? Local Law 144 applies to employees or candidates in NYC.
Was an AEDT used? Look for automated scoring, ranking, recommendation, or filtering based on algorithmic processing.
Did it “substantially assist” the decision? If the tool meaningfully influenced who advanced or was rejected, it may qualify—especially where the employer relied on the tool’s output rather than merely using it as a clerical aid.
Step 2: Capture evidence immediately (before it disappears)
AI hiring disputes are evidence-sensitive. Encourage clients to preserve:
Job posting and application pages (screenshots, PDFs, URLs, timestamps).
All automated emails and portal messages, including rejection notices and assessment invitations.
Assessment instructions and disclosures (especially any Local Law 144 notice language).
Scores, feedback reports, or “badges” the platform displays (even partial indicators can matter).
Accommodation requests and employer responses (email threads, chat logs, support tickets).
Where possible, preserve the browser’s page source or use a web capture tool. If counsel becomes involved quickly, consider sending a litigation hold letter to the employer and, where appropriate, the vendor, requesting preservation of model versions, scoring logs, audit materials, and decision rationales.
Step 3: Check the employer’s public bias audit posting
Local Law 144 requires publication of a bias audit summary. Practically, this is often found on a careers page footer, “AI notice” page, or HR compliance page. Look for:
The date of the audit (must be within the prior year before use).
The auditor identity and whether it appears independent.
Metrics indicating selection rates and impact ratios by category.
The specific tool and version audited—mismatches can be a major issue (e.g., the audit covers a résumé screener, but the employer used a video analysis tool).
Step 4: Assess notice compliance and accommodation pathways
Notice problems are common and can include:
Late notice (e.g., after the assessment is completed).
Vague notice that does not identify the job qualifications/characteristics the tool will assess.
No meaningful method to request an alternative selection process or to seek accommodation.
Where disability is implicated, evaluate whether the employer offered a reasonable alternative assessment method or adjusted the process (extended time, alternative format, human review, etc.). Under NYCHRL and ADA standards, rigid “one-size-fits-all” AI assessments can be problematic.
Step 5: Build the discrimination theory (disparate treatment, disparate impact, or failure to accommodate)
Local Law 144 violations can support the narrative that the employer used a high-risk tool without required safeguards. But the core liability often rests on:
Disparate treatment: evidence the system or its deployment treated protected groups differently (e.g., human reviewers override AI results for some candidates but not others; or the tool is tuned to disfavor certain accents or names).
Disparate impact: statistical or comparative evidence that the AEDT disproportionately screened out a protected group. This can be developed via discovery, pattern evidence, or a combination of applicant comparisons and internal hiring data.
Failure to accommodate: the strongest cases may involve a documented request for accommodation for an assessment format that penalizes disability traits.
Step 6: Choose the enforcement path
Potential paths include:
Administrative complaints (e.g., NYC Commission on Human Rights for NYCHRL; New York State Division of Human Rights for NYSHRL; EEOC for federal claims). These can be strategic for early investigation, tolling issues, and potential mediation.
Civil litigation in state or federal court, depending on claims and procedural posture. Civil discovery is often the
