This Bias, Fairness & Explainability Statement ("AI Policy") describes how ExpertsHub designs, validates, deploys, monitors, and governs AI-assisted features to minimize unlawful bias, support job-related decision making, and provide transparency to Users. It is incorporated by reference into the Terms of Use.
1. Scope and Purpose
This Statement governs AI features used for: profile and job understanding (semantic search and matching); ranking and recommendations; standardized assessments and AI-facilitated interviews; safety and fraud controls; and workflow assistance. These AI features are assistive — they generate signals to support, not replace, human judgment.
Clients remain solely responsible for hiring and engagement decisions and for compliance with employment, equal opportunity, and data protection laws in all applicable jurisdictions.
2. How Our AI Works
Our models focus on job-related textual evidence: declared and inferred skills, work history, portfolio content, certifications, role criteria, standardized questionnaire answers, and interview transcripts. We do not intentionally use protected characteristics (such as sex, gender, race, ethnicity, religion, sexual orientation, disability, or age) as model inputs, and we take steps to detect and suppress obvious proxies. Location data is used only where job-related and lawful.
We do not score or evaluate facial expressions, body language, visual backgrounds, or voice tone. Scoring is derived from content — textual resume and profile data and transcripts.
3. Bias Mitigation
We apply the following controls to reduce bias in our AI systems:
- Design controls. Feature selection reviews to reduce reliance on signals that add variance without job-relevant value.
- Pre-deployment testing. Before enabling a model, we evaluate validity and, where practicable, review group-difference indicators to detect adverse disparities and adjust as needed.
- Post-deployment monitoring. We monitor score distribution stability, completion rates, and recurring disparities. Material anomalies trigger investigation and corrective action.
- Expert review. We may engage internal or external expertise to review fairness findings as our team and resources scale.
4. Transparency and Explainability
We provide plain-language descriptions of where AI is used, what the system evaluates, and how to interpret outputs. On request, a Candidate may receive an explainability summary indicating: (a) whether sufficient content was available to score; (b) the competencies assessed; and (c) general factors that tended to influence competency indications. We do not disclose source code, weight vectors, or training data.
5. Appeals and Manual Review
Any User may request manual review of a Platform-generated signal that materially affected their experience by contacting support@expertshub.ai. We will verify configuration, evaluate logs, and re-review the AI-produced signal. We will respond within a reasonable time. Clients retain the final hiring decision.
6. Accessibility and Accommodations
Transcription accuracy may vary with accent, microphone quality, and connectivity. We implement safeguards (minimum content thresholds) and, where enabled, allow retakes or manual review. Where appropriate, configurations may support reasonable accommodations such as extended time windows or alternate formats.
7. Human Oversight
AI signals are assistive and are not a substitute for human judgment, particularly for legally significant determinations. Where applicable law restricts automated decision-making, Users must ensure AI signals are not the sole or decisive factor and must implement compliant review processes. Clients must ensure required human review and may not use AI signals to discriminate based on protected characteristics.
8. Governance
ExpertsHub oversees AI model development, testing, deployment, monitoring, and deprecation. We maintain internal documentation and change logs for our AI systems. As the team scales, we will expand our governance structure accordingly.
If we identify a material model defect or fairness risk, we may pause impacted scoring, roll back, rescore where appropriate, and notify affected Users.
9. Data and Security
We process User content to operate and improve AI features, consistent with the Privacy Policy. Where feasible, model improvements use aggregated or de-identified data. We do not use private service contract content to train public models without consent.
10. Disclaimers
No AI system is entirely free from bias or error. Our commitment is to continuously minimize and monitor risks, maintain job-relatedness, and improve fairness. ExpertsHub does not guarantee any engagement, rate, or outcome and does not warrant that recommendations or scores will be complete, timely, or error-free.
11. Contact
Explainability and Appeals: support@expertshub.ai
Abuse and AUP Concerns: info@expertshub.ai
Security Incidents: security@expertshub.ai
We may update this Statement to reflect changes in laws, standards, technology, or AI features. Updates are effective upon posting.
