Sample Bias Audit Report

Annotated example showing what you get from an OnHirely audit.

AEDT Bias Audit Report

Acme Corp — AI Resume Screening Tool

Report Date: January 15, 2025

Report ID: AUD-2025-0042

The header identifies the employer, the specific AI tool audited, and the report date. NYC LL144 requires this to be publicly posted on your website.

Executive Summary

62

Compliance Score

Moderate Risk

1,247

Candidates Analyzed

3

Adverse Impact Findings

The executive summary gives decision-makers a quick overview. Scores below 70 indicate compliance gaps that need attention.

4/5ths Rule Analysis — Gender

GroupApplicantsSelectedRateImpact RatioResult
Male68224535.9%1.00Reference
Female56514225.1%0.70 Fail

The impact ratio for Female (0.70) is below the 0.80 threshold, indicating adverse impact. This means Female candidates are selected at only 70% of the rate of Male candidates.

4/5ths Rule Analysis — Ethnicity

GroupApplicantsSelectedRateImpact RatioResult
White41215637.9%1.00Reference
Asian29810535.2%0.93 Pass
Black3128627.6%0.73 Fail
Hispanic2256227.6%0.73 Fail

Two ethnic groups (Black and Hispanic) show adverse impact with ratios below 0.80. This is a significant finding that requires immediate attention.

Statistical Significance

Gender (Female vs Male) — Chi-squared testp = 0.003 (Significant)
Ethnicity (Black vs White) — Chi-squared testp = 0.008 (Significant)
Ethnicity (Hispanic vs White) — Chi-squared testp = 0.012 (Significant)
Ethnicity (Asian vs White) — Chi-squared testp = 0.421 (Not Significant)

Statistical significance tests confirm whether the observed differences are likely due to real bias (p < 0.05) rather than random chance. Three of four comparisons are statistically significant.

Alerts & Recommendations

Critical: Gender-based adverse impact detected

Female selection rate (25.1%) is significantly lower than Male (35.9%). Review screening criteria for gender-correlated features.

Critical: Ethnicity-based adverse impact detected

Black and Hispanic groups have selection rates below the 4/5ths threshold. Investigate AI scoring model for proxy variables correlated with ethnicity.

Warning: Review candidate notification process

Ensure all candidates are notified at least 10 business days before AI screening begins, per NYC LL144 requirements.

The alerts section prioritizes findings by severity (Critical, Warning, Info) and provides actionable recommendations for each issue found.

Regulation Compliance Summary

NYC Local Law 144
Needs Action
EEOC Uniform Guidelines
Non-Compliant

The compliance summary maps audit findings to specific regulations so you know exactly where you stand and what needs to be addressed.

SHA-256 Hash: a3f2...8b4c

Generated by OnHirely

Each report includes a SHA-256 hash for tamper-proof verification. This ensures the report hasn't been modified after generation — important for regulatory submissions.

Generate your own report

Upload your hiring data and get a comprehensive bias audit report like this one in minutes. Free plan includes 1 audit with up to 200 candidates.

Generate your own report