Twenty-six Meta employees sued the company in federal court in Oakland late Monday, alleging that the AI systems Meta used to select 8,000 layoffs — about 10% of its workforce, announced in May — disproportionately picked people on medical, parental, or family leave. All 26 have received layoff notices. All 26 are, for one more week, still employees: separations begin July 22.
The complaint describes the selection machinery in unusual detail. It claims Meta fed keystroke and activity-monitoring data, AI token-usage dashboards, and algorithmically assisted performance rankings into the process that decided who stayed. The plaintiffs’ core argument is arithmetic, not ideology: many of those scores, the suit says, “by design, cannot be accumulated by an employee who is on protected medical or family leave, or whose output is reduced by a disability.” A system that ranks you by activity will always rank the person on maternity leave last — and Meta, the filing alleges, “did not pause the system for the individualized, leave- and accommodation-neutral review that the law requires.” The suit invokes the Family and Medical Leave Act, the Americans with Disabilities Act, the Pregnancy Discrimination Act, and the Pregnant Workers Fairness Act.
The plaintiff roster reads like a list of exactly the people leave laws exist to protect: eight women who took maternity or pregnancy-related leave, four men who took parental leave, one woman who took leave to care for a family member and then bereavement leave. The suit lands in a quarter where AI-attributed job cuts have gone from taboo to talking point — JPMorgan told investors this week AI has already cut some teams by 40%, and more than 200 economists just signed a public warning about AI-driven displacement. What’s new here isn’t that AI took the jobs. It’s that a court may now examine how.
Our take: The discovery fight is the real story. If this case proceeds, Meta may have to produce the actual inputs and weights behind a layoff list — the first public autopsy of algorithmic job-cutting at a major employer, and a precedent every company running “AI efficiency” layoffs has to price in. The most revealing allegation is the token-usage dashboard: companies have started grading workers on how much AI they use, and that metric quietly zeroes out anyone whose leave — or disability — keeps them away from the tools. Employment law was not written for selection-by-dashboard, but leave protections don’t care how the list got made. If layoffs-by-algorithm are entering their legal era, your own math still matters most — the Personal Runway Playbook is the 20-minute version.
What to watch
- July 22: separations are scheduled to begin. Whether Meta proceeds with the named plaintiffs’ exits or pauses them is the first signal of how it will fight this.
- Meta’s defense: contesting the claims means explaining the selection system on the record — potentially the most detailed public documentation yet of how an algorithmic layoff actually works.
- Copycat filings: 8,000 people went through the same process. More suits — or a class action — would turn one case into a category.