Human Oversight & Fairness

Our AI Recruiter is never used to make autonomous hiring decisions. All AI-generated outputs — including candidate scores, interview transcripts, and behavioral signals — are reviewed by trained human evaluators before any decision is made. This approach ensures fairness, accuracy, and full legal defensibility.

Human-in-the-Loop (HITL) by Design

Our platform is built with mandatory human checkpoints throughout the hiring workflow:

• AI scores are advisory only — They support human judgment, never replace it.
• All transcripts and recordings are reviewed before candidates are advanced or rejected.
• Bias audits are integrated into the review workflow to monitor for potential discrimination.

Every client can configure their preferred oversight model in collaboration with our team, and enterprise clients receive onboarding support to train internal reviewers.

Bias Mitigation & Fairness Checks

We implement multiple layers of bias control:

• Predefined scoring rubrics for consistency across evaluations
• Bias checklists during human reviews (race, gender, language, disability)
• Secondary reviewer escalation for flagged anomalies
• Anonymized fairness audits conducted regularly by our QA team

Our methodology aligns with EEOC guidance, GDPR Article 22, and the fairness obligations outlined in the New York AEDT Law and the proposed EU AI Act.

Client Oversight & Documentation

Clients using our AI Recruiter maintain full control over:

• Final candidate decisions
• Thresholds for score interpretation
• Custom rejection criteria
• Retention of review logs for auditability

All decisions made by reviewers — including score adjustments — are logged in a machine-readable, access-controlled audit trail for transparency and compliance.

Human Oversight Policy

HITL Oversight for Clients