Our Commitment
ZT Nation uses AI to match customers with qualified contractors. We are committed to ensuring our AI systems operate fairly and without discrimination. We monitor all 9 federal protected classes, conduct automated bias audits every 30 days, and maintain impact ratio thresholds that exceed EEOC requirements by 12.5%. Our five-layer self-healing architecture automatically detects and corrects bias drift before it can affect outcomes.
This page is published in compliance with NYC Local Law 144, Colorado SB 24-205, and EEOC guidance on AI in employment decisions.
Current AI Fairness Status
Last Audit
Pending first audit cycle
Decisions Analyzed
0
Audit Frequency
Every 30 days
ZT Nation monitors all federally protected classes in our AI scoring systems. Our AI does not collect, store, or use any demographic data. Instead, we use geographic proxy analysis to detect potential disparate impact.
Race
Title VII, Section 1981
Color
Title VII
Sex (incl. sexual orientation, gender identity)
Title VII, Bostock v. Clayton County (2020)
Religion
Title VII
National Origin
Title VII
Age (40+)
ADEA
Disability
ADA
Genetic Information
GINA
Pregnancy
PDA
We use the EEOC four-fifths rule as our baseline, with ZT Nation's premium standard set 12.5% higher. Impact ratios measure whether all groups receive proportionally equal treatment in our AI matching system.
Compliant. Documented in monthly report.
Investigation within 7 days. Root cause identified.
Root cause analysis within 48 hours. Executive notification.
Immediate system pause. AI scoring halted. Manual review required.
The EEOC four-fifths rule sets the legal threshold at 80%. ZT Nation's premium standard of 90% provides an additional 10-percentage-point buffer above the legal minimum.
ZT Nation's AI systems include automated safeguards that detect and correct bias drift without human intervention. This architecture runs continuously and is modeled after methodologies from ORCAA, Holistic AI, BABL AI, and Credo AI.
Continuous Monitoring
Every scoring decision is logged and analyzed in real-time. Impact ratios computed by geographic proxy groups every 15 minutes.
Threshold Correction
When impact ratios drop below 90%, selection thresholds are automatically adjusted per group to restore equal selection rates.
Weight Recalibration
At 88% impact ratio, scoring weights are automatically recalibrated. Maximum 3% single correction, 15% cumulative per 30-day period.
Circuit Breakers
Below 80%, AI scoring is automatically paused for the affected group. Contractors matched via FIFO queue until resolved.
Root Cause Diagnosis
Automated analysis identifies the source of drift and generates corrective recommendations with confidence scoring.
Circuit breakers automatically pause AI scoring for any group where the impact ratio falls below 80%. When tripped, contractors are matched via a first-in-first-out queue until the issue is resolved.
All AI scoring systems are operating normally within acceptable thresholds.
Continuous
Real-time monitoring of every scoring decision. Automated alerts at 88% threshold.
Monthly (30-Day)
Comprehensive statistical review with intersectional analysis. Automated report generation.
Annual Deep Audit
5-phase independent audit modeled after ORCAA, Holistic AI, BABL AI, and Credo AI methodologies.
No demographic data collected. Our AI does not ask for, store, or use race, sex, age, religion, or any other protected characteristic in scoring decisions.
Proxy variable monitoring. We actively monitor geographic and service-type variables for correlation with protected characteristics using Pearson correlation analysis.
Intersectional analysis. We test combinations of groups (not just individual categories) to detect compounded disparate impact per FTC Rite Aid consent order standards.
Scoring transparency. Contractor scores are based solely on: rating, experience, reliability, proximity, availability, skill match, and trust score. No hidden factors.
Report a Concern
If you believe you have experienced discrimination or unfair treatment through our AI matching system, we want to hear from you. All reports are reviewed within 24 hours. Anonymous reporting is available. Anti-retaliation protections apply to all reporters.
Download Anti-Retaliation Policy (PDF)Multiple reporting channels: Web form, guided chatbot, SMS (text BIAS to our number), or email. All channels support anonymous submission.
External reporting: EEOC (1-800-669-4000) · FTC · State Attorney General
This transparency report is published pursuant to NYC Local Law 144 (2021), Colorado SB 24-205 (2024), and EEOC Technical Assistance on AI and Title VII (2023).
Methodology references: EEOC four-fifths rule (29 CFR 1607.4D), Bostock v. Clayton County, 590 U.S. 644 (2020), Section 1981 of the Civil Rights Act of 1866.
Last updated: 3/16/2026 | Audit framework modeled after ORCAA, Holistic AI, BABL AI, and Credo AI