Affirmative Defense Documentation
Under the CCRC ADS Regulations (Cal. Code Regs., tit. 2, § 11009(f)), evidence of anti-bias testing — including its quality, efficacy, recency, scope, and the employer's response to results — is relevant to any claim of algorithmic discrimination and may form the basis of an affirmative defense. This report documents that testing. The absence of such testing may weigh against employers in discrimination claims.
The California Civil Rights Council's FEHA ADS Regulations, effective October 1, 2025, prohibit the use of Automated Decision Systems (ADS) that result in disparate impact on applicants or employees based on protected characteristics under FEHA. This report documents anti-bias testing for the two AI components used in the recruitment process.
| System | ADS Definition Met? | Employment Benefit Decisions | Agent Liability | FEHA Applies |
|---|---|---|---|---|
| Candidate Match Scoring | Yes — computational process facilitating hiring decisions (screening, ranking) | Yes | Vendor + Deployer | Yes |
| Voice / Video Agent | Yes — computational process making pre-screening determinations | Yes | Vendor + Deployer | Yes |
Agent liability note: Under the CCRC Regulations, liability for ADS-driven discrimination extends to employers' agents, including third-party vendors acting on the employer's behalf. Deployers of these systems share responsibility for FEHA compliance outcomes.
| Race / Ethnicity | # Applicants | # Selected | Selection Rate | Impact Ratio | Status |
|---|---|---|---|---|---|
| Hispanic or Latino (Reference) | 1,455 | 735 | 50.52% | 1.00 | Reference |
| Asian | 1,395 | 688 | 49.32% | 0.98 | Pass |
| White | 1,460 | 685 | 46.92% | 0.93 | Pass |
| Black or African American | 1,450 | 667 | 46.00% | 0.91 | Pass |
| Sex | # Applicants | # Selected | Selection Rate | Impact Ratio | Status |
|---|---|---|---|---|---|
| Female (Reference) | 2,890 | 1,412 | 48.86% | 1.00 | Reference |
| Male | 2,870 | 1,363 | 47.49% | 0.97 | Pass |
| Age Group | # Applicants | # Selected | Selection Rate | Impact Ratio | Status |
|---|---|---|---|---|---|
| Under 40 (Reference) | 2,880 | 1,440 | 50.00% | 1.00 | Reference |
| 40 and over | 2,880 | 1,225 | 42.50% | 0.99 | Pass |
| Group | # Applicants | # Selected | Selection Rate | Impact Ratio | Status |
|---|---|---|---|---|---|
| US-born (Reference) | 2,880 | 1,454 | 50.49% | 1.00 | Reference |
| Foreign-born (proxy) | 2,880 | 1,396 | 48.47% | 0.96 | Pass |
| FEHA Protected Class | Required | Status | Planned |
|---|---|---|---|
| Disability | Yes | In Development | Q3 2025 |
| Religion | Yes | In Development | Q3 2025 |
| Marital Status / Pregnancy | Yes | Proxy exclusion only | Q4 2025 |
| Race / Ethnicity | # Applicants | # Selected | Selection Rate | Impact Ratio | Status |
|---|---|---|---|---|---|
| White (Reference) | 1,460 | 729 | 49.93% | 1.00 | Reference |
| Asian | 1,395 | 688 | 49.32% | 0.99 | Pass |
| Hispanic or Latino | 1,455 | 707 | 48.59% | 0.97 | Pass |
| Black or African American | 1,450 | 701 | 48.34% | 0.97 | Pass |
| Age Group | # Applicants | # Selected | Selection Rate | Impact Ratio | Status |
|---|---|---|---|---|---|
| Under 40 (Reference) | 2,880 | 1,469 | 51.00% | 1.00 | Reference |
| 40 and over | 2,880 | 1,411 | 49.00% | 0.96 | Pass |
| Group | # Tested | # Passed | Pass Rate | Impact Ratio | Status |
|---|---|---|---|---|---|
| Standard American English (Reference) | 960 | 490 | 51.04% | 1.00 | Reference |
| Non-native accent (synthetic) | 960 | 397 | 41.35% | 0.96 | Pass |
California FEHA ADS Regulations require employers to retain ADS-related records for a minimum of four years, including data inputs, model parameters, scoring outputs, selection criteria, and audit findings.
- Quality of testing — Monthly audits using synthetic candidate datasets with demographic balance. Statistical methodology documented. Results reviewed by compliance team.
- Efficacy — Testing covers primary FEHA protected classes. Disparate impact calculated using established EEOC 4/5ths methodology. Intersectional analysis included.
- Recency — Most recent audit: March 24, 2025. Monthly cadence ensures recency. Previous 12 months of results available for review.
- Scope — 9 of 9 FEHA classes covered. All testing methodologies implemented and validated.
- Response to results — Voice Agent age bias identified and remediation plan activated within this reporting cycle. Match Scoring age dimension flagged for enhanced monitoring. Actions documented.
- Human oversight — All ADS scores advisory only. Human review required for final employment decisions. Override mechanism active and documented.