VeraHire has completed an independent, third-party bias audit of its AI hiring tools, conducted by Joellie AI, applying the standard set out in New York City’s Local Law 144 for automated employment decision tools. Reviewing real applicant data, the audit determined that VeraHire’s system demonstrated no conclusive evidence of disparate impact across race, ethnicity, sex, or their intersections across both VeraHire’s resume-screening tool and its interview analysis tool. We also tested along another vector. A preliminary counterfactual analysis of the resume screener also found no material bias in how it scores otherwise-identical applications.
Why did we do this audit?
VeraHire scores each candidate against the specific requirements in a job description, one criterion at a time. Every score has to trace back to something the candidate either wrote or said.
We asked a certified, independent bias audit firm to analyze how our tools score real candidates, using the standards regulators and courts apply to employment decisions. We want to enable hiring teams to make consistent, evidence-based decisions, and ensure every candidate gets an equivalent opportunity.
What did the audit examine?
The core audit measured disparate impact: whether the scores led to different outcomes for candidates from different backgrounds, judged against the four-fifths rule used under NYC Local Law 144 and in federal Title VII enforcement. The auditor worked from actual anonymized resume-screening data and first-round interview transcripts, across race, ethnicity, sex, and their intersections. The audit determined VeraHire systems contain no conclusive evidence of disparate impact in either tool.
As a preliminary further step, to be sure, we then asked the auditor to run a counterfactual “name-swap” analysis of our AI resume screener. Holding each resume’s qualifications constant, the auditor changed only the applicant’s name to test any outcome score variance that might have been triggered by an inherent AI association related to certain names or name types rather than by the candidate’s record. These early results also found no material bias: the name made no meaningful difference to the score or system assessment.
We asked for that extra analysis because the hiring teams who rely on VeraHire carry real exposure on fairness, compliance and legal risk. An independent review gives them something to stand on. A candidate’s name should have no bearing on a job application evaluation. We wanted a qualified independent audit firm to analyze, rigorously test, and confirm this.
Does the audit cover everything on the platform?
The auditor focused on VeraHire’s core candidate assessment tools. We want to clarify that this evaluation is not a formal certification of VeraHire overall, the full extent of the product platform, or the underlying artificial intelligence models VeraHire uses, and it is not a guarantee that the hiring process is free of bias. The results are specific to these identified tools, system versions, and data examined. In all situations with use of the VeraHire platform, a human always still makes every hiring decision.
Our solid and continuous commitment to fairness
One audit does not settle the bias question on an ongoing basis. We are committed to continued and rigorous objective anti-bias assessment and review and will maintain regular, periodic system audits going forward.
The full results, including selection rates and impact ratios for every category, are in the bias audit summary.
Learn more about VeraHire’s approach to structured hiring at verahire.ai.
About VeraHire
VeraHire is an AI recruiting platform that screens resumes and runs structured first-round interviews, so hiring teams move faster and every candidate gets the same first round.
Press contact: press@verahire.ai