Skip to main content
✅ PRODUCTION ENGINE

NYC Hiring Law Bias Detection

Production-grade compliance analysis for NYC Local Law 144 and federal anti-discrimination laws

✓ Local Law 144 Compliance
✓ AEDT Bias Audit
✓ Title VII / ADA / ADEA
✓ Fair Chance Act

Bias Detection Engine — Pro Feature

The NYC LL144 Bias Detection Engine is available on the Pro plan and above. Analyze job descriptions, screening criteria, and hiring decisions for compliance with NYC Local Law 144, Title VII, ADA, ADEA, and the Fair Chance Act.

Upgrade to Pro — $299/moLearn About NYC LL144 Compliance

Pro includes: Bias Detection Engine, 3 team seats, unlimited compliance roadmaps, API access, and more.

What is NYC Local Law 144?

Enacted in 2021, NYC Local Law 144 requires employers and employment agencies to:

  • •Conduct annual independent bias audits of AEDTs
  • •Publish audit results publicly
  • •Notify candidates when AEDTs are used
  • •Provide alternative selection process upon request

How This Tool Helps

Our bias detection engine analyzes your hiring materials for:

  • ✓Gender-coded language and pronouns
  • ✓Age bias indicators
  • ✓Race and ethnicity coded terms
  • ✓Disability discrimination
  • ✓Criminal history violations (Fair Chance Act)
  • ✓Geographic and educational barriers

Need Professional Bias Audit Services?

While this tool provides preliminary screening, NYC Local Law 144 requires independent bias audits by qualified third parties. HAIEC offers comprehensive AEDT auditing services.

Example Analysis

See how the tool analyzes job posting language for bias indicators:

Input Text:

"Looking for a rockstar developer who can work in a fast-paced, high-energy environment. Must be a cultural fit and able to work long hours when needed."

Analysis Result:

⚠️ Bias Risk: HIGH

Multiple bias indicators detected:

  • "Rockstar" - May exclude candidates who don't identify with aggressive, masculine-coded language
  • "Cultural fit" - Vague criterion often used to exclude protected groups
  • "Long hours" - May discriminate against candidates with caregiving responsibilities (disparate impact on women)

Better Alternative:

"Seeking a skilled software engineer with 3+ years of experience in [specific technologies]. Must meet project deadlines and collaborate effectively with distributed teams."

✓ Bias Risk: LOW

Uses objective, measurable criteria without coded language or vague cultural requirements.

About the Author

Subodh KC is the founder of HAIEC and author of the Instruction Stack Audit Framework (ISAF). His research on AI accountability has been published on Zenodo (DOI: 10.5281/zenodo.14555643).

Last reviewed: January 2026

How NYC Enforces Local Law 144

Enforcement Authority and Mechanism

Primary Enforcer: NYC Department of Consumer and Worker Protection (DCWP)

The DCWP's Office of Labor Policy & Standards oversees LL144 enforcement through:

1. Complaint-Driven Investigations

  • Job candidates can file complaints online or by phone
  • Current employees can report violations
  • Complaints trigger formal investigations within 30 days
  • DCWP has subpoena power for employer records

2. Proactive Compliance Audits

  • DCWP monitors public job postings for AI screening disclosures
  • Routine audits of high-risk sectors (finance, tech, healthcare, retail)
  • Cross-referencing bias audit publication requirements
  • Checking careers pages for required summaries

What Triggers an Investigation

High-Priority Triggers:

  1. Candidate Complaint - Most common trigger. Candidate suspects AI screening without notice.
  2. Missing Public Summary - DCWP monitors careers pages. Missing summary = immediate red flag.
  3. Employee Report - Current employees can report internally used AEDT with whistleblower protections.
  4. Media Reports - News coverage or viral social media complaints trigger investigations.
  5. Sector Sweeps - Industry-wide audits of all major employers in sector.

Investigation Process Timeline

Day 1-30: Initial Investigation

  • DCWP reviews complaint and public information
  • Checks careers page for bias audit summary
  • Reviews job postings for candidate notice
  • Determines if violation appears to exist

Day 31-60: Document Request

  • Formal notice of investigation sent to employer
  • Document production request (bias audit, policies, records)
  • 15-day response deadline
  • Failure to respond = separate violation

Day 91-120: Resolution

  • Option A: Warning Letter (first-time violations) - 30-day cure period, no penalty if cured
  • Option B: Notice of Violation - Penalty assessment, settlement negotiation, consent decree
  • Option C: Administrative Hearing - Employer contests violation, formal hearing

Real Violation Examples

Example 1: Missing Bias Audit Summary

Company: Mid-size financial services firm (1,200 employees)

Violation: Used AI resume screening for 8 months without publishing bias audit summary

Discovery: Candidate complaint after rejection

Investigation findings:

  • Bias audit existed but wasn't published
  • Careers page had no mention of AEDT use
  • Candidate notice was missing

Outcome: Warning letter (first-time violation), 30-day cure period granted, published bias audit summary, added candidate notice. No financial penalty.

Example 2: No Bias Audit Conducted

Company: Tech startup (300 employees)

Violation: Used AI screening for 12 months without any bias audit

Discovery: DCWP proactive audit of tech sector

Investigation findings:

  • No bias audit ever conducted
  • Assumed vendor's audit satisfied requirement (it didn't)
  • No independent auditor engaged
  • No candidate notice provided

Outcome: Notice of violation. 365 days × $500 = $182,500 minimum exposure. Settlement: $125,000 + mandatory independent audit + 2-year monitoring + quarterly compliance reports.

Example 3: Inadequate Candidate Notice

Company: Healthcare system (5,000 employees)

Violation: Notified candidates only 3 days before AEDT use (law requires 10 days)

Investigation findings:

  • Notice timing violated 10-day requirement
  • Notice language was vague about AI's role
  • No opt-out or alternative process offered

Outcome: 180 days × $1,500 = $270,000 maximum exposure. Settlement: $175,000 + revised notice process (10+ days advance) + clear explanation of AI's role + alternative application process.

What Makes a Valid Bias Audit Under LL144

Statutory Requirements

1. Independent Auditor

  • No financial interest in employer or AEDT vendor
  • Cannot be employee of employer or vendor
  • Cannot have business relationship beyond audit engagement
  • Must have relevant expertise (statistics, employment law, AI)

What counts as "independent": Third-party consulting firm, academic researcher, specialized bias audit firm, employment law firm with statistical expertise

What doesn't count: Vendor's internal audit team, employer's internal data science team, consultant with ongoing advisory relationship, auditor with equity stake in vendor

2. Bias Metrics: Impact Ratios

Must calculate selection rates and impact ratios for:

  • Sex: Male vs. Female
  • Race/Ethnicity: Per EEOC categories (White, Black/African American, Hispanic/Latino, Asian, Native American, Two or More Races, Other)
  • Intersectional categories: Sex × Race/Ethnicity combinations

Impact ratio formula:

Impact Ratio = (Selection Rate for Category) / (Selection Rate for Most Favored Category)

Example:
- Female selection rate: 15%
- Male selection rate: 20%
- Impact ratio for females: 15% / 20% = 0.75

Disparate impact threshold: Impact ratio < 0.80 suggests disparate impact (based on EEOC's "four-fifths rule"). Not automatic violation, but requires justification.

3. Data Requirements

  • Minimum sample size: At least 100 individuals per category (if available)
  • Data recency: Most recent 12 months of AEDT use
  • Data completeness: All individuals screened by AEDT, cannot cherry-pick favorable time periods

4. Publication Requirements

What must be published:

  • Summary of bias audit results
  • Impact ratios for all tested categories
  • Date of audit
  • Auditor information (name, firm)
  • AEDT vendor information

Where to publish: Employer's careers page or website, accessible to public without login, available for at least 6 months after audit date

5. Annual Update Requirement

  • New audit required every 12 months
  • Triggered by material changes to AEDT
  • Must reflect current system version

Frequently Asked Questions

What is an AEDT under NYC Local Law 144?

An Automated Employment Decision Tool (AEDT) is any computational process, derived from machine learning, statistical modeling, data analytics, or artificial intelligence, that issues simplified output (score, classification, or recommendation) used to substantially assist or replace discretionary decision-making for employment decisions. This includes AI resume screening, automated video interview analysis, predictive hiring assessments, algorithmic candidate ranking, and AI-powered skills testing.

Does LL144 apply to AI scheduling or time tracking tools?

Generally no. LL144 specifically covers 'employment decisions' defined as hiring, promotion, or other employment opportunities. Scheduling and time tracking are operational, not decisional. However, exceptions exist: if shift assignments influence promotion eligibility, if scheduling impacts performance reviews, or if time tracking data feeds into promotion decisions, then it may be considered AEDT.

What if my bias audit shows disparate impact?

You can still use the AEDT under LL144. Having disparate impact doesn't prohibit AEDT use—you must publish results (including disparate impact findings) and can continue using the tool. However, consider Title VII risk: disparate impact can trigger federal discrimination claims. You must prove business necessity and show no less discriminatory alternative exists. If audit shows significant bias (impact ratio < 0.60), strongly consider remediation before continued use.

Do I need a new bias audit if I change AI vendors?

Yes, in most cases. New vendor = new AEDT with different algorithms, training data, and decision logic. Previous audit doesn't cover new system. New audit required when: switching to different vendor's AEDT, upgrading to new model version from same vendor, material changes to AEDT configuration, or adding new features. Plan 4-8 weeks for audit completion. Cannot use new AEDT while waiting for audit.

How do I find an independent auditor for LL144 compliance?

Look for auditors with: statistical expertise (impact ratio calculation), employment law knowledge (protected categories, EEOC standards), AI/ML understanding, and no financial interest in employer or vendor. Sources: specialized bias audit firms ($15K-$50K), employment law firms with statisticians ($25K-$75K), academic researchers ($10K-$30K), or HR consulting firms ($20K-$60K). Red flags: auditor has ongoing advisory relationship with employer, affiliated with AEDT vendor, has equity stake in vendor, or fee contingent on favorable results.

Can I use the same bias audit for multiple job roles?

It depends on how the AEDT is used. One audit sufficient if: same AEDT used across all roles, same decision criteria applied, same screening process, and audit covers all roles collectively. Separate audits required if: different AEDTs for different roles, different decision thresholds by role, role-specific screening criteria, or material differences in application. Best practice: conduct comprehensive audit covering all AEDT uses and document which roles are covered.

What happens if I'm investigated but disagree with DCWP's findings?

Dispute process: (1) Informal resolution (Days 1-30): respond to initial findings, provide additional documentation, request meeting. (2) Formal response (Days 31-60): submit written response to Notice of Violation, cite legal authorities. (3) Settlement negotiation (Days 61-90): negotiate penalty amount, agree on remediation. (4) Administrative hearing (Days 91+): present evidence and witnesses, cross-examine DCWP witnesses. (5) Appeal if adverse decision. Legal fees for hearing/appeal: $50K-$200K+. Settlement often more cost-effective.