Insights - AI Bias Audit Report

The AI Bias Audit Report helps organizations monitor the fairness of AI-powered candidate recommendations by providing visibility into how recommendations are distributed across different demographic groups. While AI Candidate Match helps recruiters quickly identify qualified applicants, the report ensures those recommendations remain transparent and can be reviewed for potential bias. AI recommendations are intended to support hiring decisions—not replace them—and final decisions should always remain with your recruiting team.

🎥 The following video provides an overview of this feature:

How It Works

AI Candidate Match evaluates each applicant’s resume against your specified hiring criteria, including Must Have, Ideal, and Nice to Have qualifications. These criteria are generated from the job description but can be reviewed and adjusted by recruiters before candidates are evaluated. Based on this comparison, candidates are categorized as Recommended, Review, or Not Recommended.

The AI Bias Audit Report analyzes these recommendation outcomes to identify whether recommendation rates differ significantly across demographic groups. To access the report, navigate to Insights → AI Bias Audit Report, where you can review data for a selected date range and filter results by department, job, or location.

Understanding the Report

The report compares recommendation rates across demographic categories such as gender, ethnicity, veteran status, disability status, and other fields available within your ATS. Importantly, this information is never inferred by AI. All demographic data comes directly from candidate self-identification during the application process, ensuring the analysis is based only on information voluntarily provided by applicants.

For each demographic category, the report establishes a baseline using the group with the highest recommendation rate. Other groups are then compared against that baseline to calculate potential disparate impact. For organizations hiring in the United States, the report also helps monitor alignment with New York City Local Law 144, highlighting whether recommendation rates remain within the commonly referenced 80% threshold.

If there is insufficient data to produce statistically meaningful comparisons, the report displays a warning indicator, helping prevent conclusions from being drawn from very small sample sizes.

Global Support

Because hiring practices and demographic reporting requirements vary around the world, the report automatically adapts based on the country where a job is posted. For example, the demographic categories available for a role in Ireland will differ from those shown for a role in the United Kingdom or the United States. These categories are driven by the data available in your ATS, allowing organizations to monitor fair hiring practices using country-specific demographic reporting.

Taking Action

If the report identifies potential disparities, the first step is typically to review the job’s AI matching criteria. Overly restrictive or unintentionally biased requirements can influence recommendation outcomes. Because recruiters can edit the AI-generated Must Have, Ideal, and Nice to Have qualifications, organizations can refine these requirements and continually improve the fairness and consistency of future AI recommendations.

The AI Bias Audit Report is available for organizations using Job Pipeline with Workday or Greenhouse. If you’d like to enable this feature, please contact early@goodtime.io, and our team will help you get started.