AI for hiring, held to a higher bar
Hiring decisions change lives. VIVA360 is designed so every model output is explainable, every reviewer is accountable, and every customer can defend their process to candidates and regulators.
Our principles
Fairness
We continuously monitor selection rates by gender, ethnicity, age, and disability where lawful. Quarterly bias audits are published to customers and aligned to NYC Local Law 144.
Transparency
Every score is accompanied by the evidence behind it. Candidates and reviewers see what was evaluated, why, and how it was weighted.
Human oversight
AI ranks and explains — humans decide. No automated rejection is sent without a recruiter-in-the-loop review on configurable thresholds.
Auditability
Immutable logs of every model decision, prompt version, and reviewer action. Exportable for regulators and internal audit.
Risk classification
We treat hiring use cases as high-risk under the EU AI Act and apply the corresponding documentation, testing, and post-market monitoring.
Continuous evaluation
Models are re-evaluated on fresh, real-world data before every release. Drift, fairness, and quality regressions block deploys.
How we govern AI
Use-case approval. New AI features are reviewed by an internal Responsible AI Council that includes engineering, legal, and external advisors. High-risk features require fairness testing, red-teaming, and a documented rollback plan before launch.
Data minimization. We collect only what is needed for the hiring decision. Protected attributes are never used as model inputs; they are used only to measure adverse impact, in line with EEOC and ICO guidance.
Candidate rights. Candidates receive notice of AI use, can request human re-review, and can access the evidence behind any decision — including the right to correct factual inputs.
Vendor LLMs. When we use third-party foundation models, customer data is never used for vendor training, and we enforce contractual zero-retention where the vendor supports it.
Reporting concerns
If you believe a VIVA360 output is biased, incorrect, or misused, email responsible-ai@viva360.ai. We acknowledge reports within 2 business days and publish a summary of material findings in our annual transparency report.