The Employment Lawyer Ensuring AI Doesn't Replicate Workplace Bias

By Sarah Hayes, JD — 2025-12-10

Sarah Hayes has spent fourteen years advising employers on discrimination claims, NLRA compliance, and wrongful termination litigation at a national employment law practice. She joined IXO because she recognized that AI was creating an employment law problem that most clients didn't yet understand they had.

AI across the HR lifecycle

"Employment AI is proliferating across the HR lifecycle," she says. "Resume screening, interview assessment, performance evaluation, compensation benchmarking, RIF selection. In each of these contexts, the use of AI creates potential disparate impact liability that most employers haven't thought through. And the legal analysis of that liability is changing fast — EEOC guidance, state AI employment laws, emerging case law."

Her IXO work focuses primarily on evaluating AI legal research and compliance tools for accuracy in employment law contexts — specifically testing whether AI tools correctly identify the applicable anti-discrimination framework, accurately characterize relevant EEOC guidance and judicial precedent, and surface the state-level variations that are increasingly important as several states enact specific AI employment legislation.

The adverse impact analysis gap

"The adverse impact analysis is one area I focus on heavily. AI tools that advise employers on selection procedure validation are surprisingly inconsistent on how to apply the standard, how to document validation efforts, and when validation is legally required versus merely advisable. Those distinctions matter enormously in litigation."

Documentation and evidence

She also evaluates AI tools in the termination documentation context. "Documentation created with AI assistance has evidentiary implications that most people haven't thought about. An AI that helps managers write performance reviews has just become part of the evidentiary record in any future wrongful termination claim."

Read The Employment Lawyer Ensuring AI Doesn't Replicate Workplace Bias on the IXO blog