Vervoe Bias Audit

Employers and employment agencies using automated employment decision tools to assess candidates or employees seeking positions in New York City must have a bias audit conducted by an independent auditor. This requirement took effect on 5 July 2023.

An automated employment decision tool is defined as any computational process derived from machine learning, statistical modeling, data analytics, or artificial intelligence that produces a simplified output - including a score, classification, or recommendation - that is used to substantially assist or replace discretionary decision making in employment decisions.

A bias audit is an impartial evaluation by an independent auditor, assessing the disparate impact of a tool against category 1 protected characteristics - specifically ethnicity/race and sex.

A summary of the results of a current bias audit (conducted within the last year) must be published on the employer’s or employment agency’s website before using the tool with candidates or employees, beginning 5 July 2023.

Vervoe has engaged Holistic AI - an independent, accredited AI auditing firm - to conduct a comprehensive bias audit of our AI scoring models. The audit assessed whether Vervoe’s assessment scoring produces disparate impact across ethnicity/race and sex categories, in accordance with the requirements of NYC Local Law 144.

Holistic AI certified Vervoe’s AI scoring as bias-free and responsible. Our models evaluate candidates solely on their demonstrated performance on job-relevant tasks - without access to demographic information, educational credentials, or resume proxies.

The audit is conducted on an ongoing basis to ensure continued compliance as our models are updated and improved.

Vervoe’s AI scoring models are trained on millions of human-graded assessment responses and calibrated against real hire outcomes. Scoring is based entirely on what candidates demonstrate in response to job-relevant tasks - written answers, video responses, coding challenges, and practical simulations.

Because our models do not receive or process demographic data, they assess candidates on merit alone - helping employers identify the best person for the job regardless of background. Every AI score is explainable: recruiters can see exactly which evidence produced each score and override the AI at any point.

Vervoe is also designed with EU AI Act requirements in mind, and our platform includes human-in-the-loop controls throughout the hiring workflow.

For questions about Vervoe’s bias audit, AI practices, or LL 144 compliance, please contact us.

For media and research inquiries: media@vervoe.com