A company used AI to analyze employee performance and identify layoff candidates.
The AI recommended 150 employees across 12 locations based on productivity scores, skill gaps, and future workforce needs.
Management implemented the layoffs. No WARN Act notice. They didn’t think it was required because no single location hit 50 employees.
Then the lawsuits came.
The WARN Act aggregates terminations across locations when they’re part of a single employment action. The AI-driven workforce reduction was one action affecting 150 people.
The company owes 60 days’ back pay and benefits to all 150 terminated employees.
How Companies Use AI for Layoffs
Workforce reductions in 2025-2026 are increasingly AI-driven.
Performance analysis: AI analyzes productivity data, performance reviews, and work output to score employees.
Skill gap identification: AI compares employee skills to future needs and identifies people with obsolete skills.
Redundancy detection: AI identifies overlapping roles and recommends consolidation.
Impact modeling: AI predicts the business impact of different layoff scenarios.
Management trusts AI analysis because it seems objective and data-driven.
The legal problems emerge when companies don’t understand how AI-driven layoffs trigger employment laws.
The WARN Act Basics
The Worker Adjustment and Retraining Notification (WARN) Act requires 60 days’ advance notice before mass layoffs or plant closings.
It applies when 50+ employees at a single site are laid off, or 500+ employees are laid off regardless of location, or 33%+ of the workforce at a site (minimum 50 employees) is laid off.
Companies must provide written notice to affected employees, the state dislocated worker unit, and the local government.
Failure to provide notice means the company owes affected employees 60 days’ back pay, benefits, and potentially penalties.
How AI Layoffs Trigger WARN
Companies using AI to select layoff targets often violate WARN without realizing it.
Aggregation across locations: AI recommends employees across multiple sites. The company doesn’t realize that the total triggers WARN even if no single site hits the threshold.
Single employment action: AI-driven layoffs over 90 days count as one action if they’re part of a single reduction plan.
Hidden patterns: AI selection criteria create patterns that show a coordinated reduction, even if the company claims individual terminations.
Timing: AI implements layoffs over weeks without a 60-day notice because the company thinks they’re separate decisions.
The AI optimizes for business outcomes. It doesn’t understand WARN Act compliance.
The Discrimination Problem
AI selecting layoff targets creates additional legal risk beyond WARN.
Disparate impact: If AI disproportionately selects older workers, minorities, or protected groups, it creates discrimination claims.
Proxy discrimination: AI uses factors that correlate with protected characteristics (like certain skill sets correlating with age).
Validation requirements: If AI selection is challenged, the company must prove the AI criteria are job-related and consistent with business necessity.
Documentation gaps: Companies can’t explain how AI made selection decisions, which makes defending layoffs nearly impossible.
Many companies discover these problems only after the layoffs, when employees sue and ask for the AI selection criteria.
State WARN Act Equivalents
The Federal WARN Act isn’t the only issue.
Many states have their own WARN-type laws with different thresholds, including California (75+ employees triggers notice), New York (25+ employees at a single site), and Illinois (25+ employees triggers notice).
Companies using AI for layoffs must comply with federal WARN and all applicable state laws.
AI-optimized layoff plans often violate multiple jurisdictions’ laws simultaneously.
What Companies Should Do
If companies are using AI to inform layoff decisions, legal review is essential before implementation.
WARN Act analysis: Calculate whether proposed layoffs trigger federal or state WARN requirements.
Adverse impact testing: Analyze whether AI recommendations disproportionately affect protected groups.
Validation review: Confirm AI selection criteria are job-related and defensible.
Documentation: Preserve records showing how AI made recommendations and what human review occurred.
Alternative scenarios: Consider whether different layoff structures avoid WARN triggers or discrimination risks.
Notice timing: If WARN applies, provide the required 60-day notice before layoffs.
Most companies implement AI recommendations without this legal review because they trust the AI’s business logic.
Most companies implement AI recommendations without legal review because they trust the AI’s business logic.
The AI optimizes for cost savings and business efficiency. It doesn’t know about WARN Act aggregation rules, state notice requirements, or adverse impact testing.
When the lawsuit comes, “the AI recommended it” isn’t a defense. The company is liable for every employment law the AI-driven layoff violated.
And 60 days of back pay plus benefits for 150 employees is just the beginning.
Brian Alexander
Founding Partner, VisionAI+ Consulting Group
Former General Counsel, AI Engineering Firm
Certified ISO 42001 Lead Auditor