HB HireBias Audit

AI hiring bias audit

AI Hiring Bias Audit Workflow

An AI hiring bias audit answers whether automated sourcing, resume screening, interview scoring, or assessment tools create materially different outcomes for protected groups. The useful output is not only a score. It is an evidence trail showing which tool, stage, group, and threshold needs review.

Run audit preview

Where this fits

A mid-market employer uses AI resume screening and needs a defensible annual review before renewing the vendor.

HR compliance has applicant-flow exports but no repeatable way to calculate adverse impact across groups.

A legal or people team needs a report that can be shared with leadership, auditors, or outside counsel.

Operating steps

  1. Inventory every AI hiring tool, vendor, stage, model version, and decision threshold used in the hiring process.
  2. Load applicant and selection counts by protected group, role family, location, and hiring stage.
  3. Calculate selection rates, impact ratios, and 4/5 rule flags for every meaningful comparison group.
  4. Review high-risk combinations with sample-size notes, human override points, and remediation recommendations.
  5. Export an audit packet with methodology, findings, limitations, and follow-up monitoring dates.

Common risks

  • Only testing the final hire decision while the AI screen creates the earlier exclusion point.
  • Combining roles or locations so aggressively that stage-specific adverse impact disappears.
  • Treating vendor claims as enough evidence without testing the employer-specific applicant flow.

How HireBias Audit connects

HireBias Audit turns the inventory, 4/5 rule calculation, risk scoring, and report draft into one workflow so HR and legal teams can move from raw counts to an audit-ready packet.

View plans