NYC Local Law 144 — AI Hiring Bias Audit Disclosure

Last updated: August 23, 2026 · Callme Group LLC

1. Overview

Callme AI's 5-dimension AI Interview scoring system is classified as an Automated Employment Decision Tool (AEDT) under NYC Administrative Code §20-874 (Local Law 144 of 2023). This page provides the required bias audit summary, scoring methodology, and candidate rights information.

2. Tool Description

The Callme AI Interview evaluates caregiver candidates across 5 dimensions:

  • Safety & Child/Senior Protection (30% weight) — Emergency response, CPR knowledge, hazard identification
  • Technical Competence (20% weight) — Age-specific care, medical conditions, daily routines
  • Communication (20% weight) — Clarity, professionalism, family communication
  • Problem-Solving (15% weight) — Scenario response, adaptability, conflict de-escalation
  • Reliability (15% weight) — Attendance, punctuality, commitment signals

The system evaluates speech content and textual reasoning only. No facial biometric analysis, facial-expression recognition, or voice biometric identification is performed.

3. Bias Audit Summary

An independent bias audit was conducted covering EEOC protected classes: race/ethnicity, gender, age 40+, disability, and national origin. The audit examined:

  1. Selection rates — Pass rate (overall score ≥ 70) by demographic group
  2. Scoring distribution — Mean scores by group per dimension
  3. Impact ratio — Ratio of the lowest-scoring group's pass rate to the highest-scoring group's pass rate (4/5ths rule)

The full bias audit report is available upon request at compliance@huwo.ai. Key findings and mitigation measures are described below.

4. Mitigation Measures

  • Temperature calibration — LLM inference temperature set to 0.3 for consistency
  • Adversarial re-scoring — 10% of interviews are re-scored by a second model instance for drift detection
  • Prompt review — Quarterly review of all interview prompts for language bias
  • Human override — Any candidate may request human-only review at no penalty
  • No biometric analysis — Eliminates facial-expression and voice-biometric bias vectors entirely

5. Candidate Rights

  • Right to request human-only review of the interview at any stage.
  • Right to receive the scoring breakdown (5-dimension scores) upon request.
  • Right to appeal the score and have a second evaluator review the interview.
  • Right to delete the interview recording and transcript at any time.

To exercise any of these rights, email compliance@huwo.ai with the subject line "NYC LL144 Request". Response SLA: 48 hours.

6. Effective Date

This disclosure applies to all AI interviews conducted on or after September 1, 2026 for candidates residing in or applying for positions in New York City (5 boroughs: Manhattan, Brooklyn, Queens, Bronx, Staten Island).

7. Authority & References

  • NYC Administrative Code §20-874 (Local Law 144)
  • DCWP Automated Employment Decision Tool Law
  • EEOC Uniform Guidelines on Employee Selection Procedures (29 C.F.R. §1607)