Chauniqua Young is an employment attorney and partner at Outten & Golden LLP in New York City, where she has spent more than a decade representing workers in disputes over unpaid wages, pay equity, and workplace discrimination. She began her legal career as a Bertha Justice Fellow at the Center for Constitutional Rights, litigating civil rights cases before joining Outten & Golden as an associate in 2014 and becoming a partner in 2021. Young earned her bachelor’s degree at Sarah Lawrence College and her JD at the Benjamin N. Cardozo School of Law, later clerking in the U.S. District Court for the Southern District of New York. She has been named among The Best Lawyers in America and the Lawdragon 500 Leading Plaintiff Lawyers, and her litigation record includes a $26 million settlement for underpaid employees. That record shapes how she evaluates whether a completed bias audit truly settles questions of hiring fairness.
Employers may use software in recruiting, screening, hiring, or promotion. Some systems screen resumes for keywords or experience, while others evaluate recorded interviews.
In New York City, a covered tool must undergo a bias audit before use. That review measures tool outcomes, but it does not determine whether the entire process is fair.
An automated employment decision tool (AEDT) is a computational process that uses machine learning, statistical modeling, data analytics, or artificial intelligence to issue a simplified output. That output may be a score, tag, classification, rank, or recommendation. Coverage turns on whether the output substantially assists or replaces discretion when employers screen job candidates or employees for promotion.
A bias audit is an impartial evaluation by an independent auditor. It tests an AEDT for disparate impact, or uneven outcomes, in specified reporting categories. It checks measured outcomes, not every part of a hiring or promotion decision.
The audit measures selection rates and impact ratios, which compare outcomes across groups, including sex, race and ethnicity, and intersectional categories. For scoring tools, it measures scoring rates. The public summary must identify where the data came from, explain the data used, and report the required category results.
What the audit can show depends on the data behind it. The review uses historical data from prior use of the tool, or test data when there is not enough historical data for a statistically significant audit. When an audit uses test data, the summary must explain why and how the auditor obtained or generated it.
Even a completed audit does not replace the required notice. At least 10 business days before use, the employer must notify a city-resident candidate or employee that the tool will be part of the process. The notice must identify the job qualifications and characteristics the tool will consider. It must also allow a candidate to request an alternative selection process or accommodation.
A score or recommendation raises another question the audit does not settle – what is the tool measuring? A system may use qualifications, experience, or recorded interview material to produce that output.
Hiring tests and screening procedures matter when they exclude applicants or employees. If a test screens out protected groups, the employer may need to show that such screening is job-related and consistent with business necessity.
The weight an employer gives the output can change the tool’s role in the selection process. An AEDT may substantially assist decision-making when an employer relies only on its output, gives that output more weight than any other criterion, or uses it to overrule other factors, including human judgment. The issue is not just the output, but the employer’s reliance on it.
Federal antidiscrimination laws still apply when employers use automated systems. A tool does not sit outside workplace law because it relies on data, scoring, or artificial intelligence. A neutral policy or practice can raise legal concerns when it disproportionately excludes a protected group and is not job-related and in line with the needs of the business.
A bias audit has a one-year limit. The AEDT rules restrict use or continued use when more than one year has passed since the most recent bias audit. Continued use remains tied to the reviewed tool, the audit data, the public summary, and the required notice.
A completed audit can mark a checkpoint, not the end of review. It can show how a covered tool performed under measured conditions, but it does not decide the fairness of the entire selection process. The remaining question is whether the employer gave notice, used job-related criteria, relied on the output within the covered screening process, and kept the audit current.
About Chauniqua Young
Chauniqua Young is a partner at Outten & Golden LLP in New York City, where she represents employees in matters involving wage violations, pay equity, and workplace discrimination. She began her career as a Bertha Justice Fellow with the Center for Constitutional Rights before joining Outten & Golden in 2014. A graduate of Sarah Lawrence College and the Benjamin N. Cardozo School of Law, she has been named to The Best Lawyers in America and the Lawdragon 500, and she serves on the board of the Public Justice law center.
