Meta Workers Sue Over AI Layoffs That Targeted Medical Leave

Twenty-six current and former Meta employees sued the company Monday, claiming the software it used to help pick who lost their jobs penalized workers on medical, parental and family leave. The suit, filed in the U.S. District Court for the Northern District of California in Oakland, targets the AI-driven scoring tools Meta leaned on during its May layoffs, arguing the math couldn’t help but punish anyone who had been away from a keyboard for a protected reason.

Meta says human managers made the layoff decisions. The plaintiffs want a judge to freeze their terminations, set for July 22, before that claim gets tested.

Inside the AI Systems That Scored Meta’s Workforce

The lawsuit describes a constellation of internal artificial intelligence systems that fed into Meta’s termination list. According to the complaint, the company drew on internal AI assistants, activity-monitoring software and algorithmically assisted performance rankings when scoring jobs for elimination this spring.

• Metamate – Meta’s internal AI assistant, cited in the complaint as one input into termination decisions.
• “Second brain” agents – AI tools trained on individual employee files and work history, according to the lawsuit.
• Productivity scoring – a system built on keystroke and screen-activity data, used to rate output.
• AI token-usage dashboards – trackers measuring how heavily employees used Meta’s own AI tools, folded into performance rankings.

The plaintiffs argue the design itself was the problem. Their filing states that several of these inputs “by design, cannot be accumulated by an employee who is on protected medical or family leave, or whose output is reduced by a disability.” Meta did not adjust scores for employees on approved leave, the complaint says, and did not pause the system for the individualized review the law requires before finalizing the cuts.

The 26 Plaintiffs, by the Numbers

The plaintiffs come from six states, including California and New York, plus Washington, D.C. They were among the roughly 8,000 employees, about 10% of Meta’s staff, whom the company said in May would lose their jobs as it cut roughly 8,000 jobs while pivoting deeper into AI.

About half the plaintiffs, in other words, took leave for pregnancy or caregiving reasons before landing on the termination list. The complaint argues the scoring process fell more heavily on women than men, since women disproportionately take pregnancy and caregiving leave under Meta’s own policies.

A First Test for New Algorithmic Bias Laws

Meta’s accusers say the company never tested its layoff tools for discriminatory impact, a step now required under recently adopted bias-audit laws in California and New York City. Those rules require employers to test automated employment decision tools for skewed outcomes before using them to hire, promote or terminate staff.

The case appears to be among the first of its kind against a major U.S. employer challenging how AI shaped layoff decisions. Economists have already flagged younger workers as the first to feel AI’s labor-market squeeze, and this case extends that scrutiny to the software companies use to decide who stays employed.

What we know:

• The 26 plaintiffs filed suit Monday, July 13, in federal court in Oakland.
• Their terminations are scheduled to take effect July 22.
• Meta attributes the layoff decisions to human managers.

What’s unconfirmed:

• Whether the court will grant a preliminary injunction blocking the July 22 terminations.
• How many of the roughly 8,000 laid-off workers beyond these 26 might bring similar claims.
• Whether Meta’s internal AI systems will face the independent audit the plaintiffs are requesting.

Meta’s AI Spending Collides with the Layoff Math

The cuts landed as Meta poured money into artificial intelligence. The company has guided to $115 billion to $135 billion in 2026 capital spending, up from $72.2 billion in 2025, driven largely by its Meta Superintelligence Labs effort. At the end of 2025, the company employed 78,865 people globally, according to its 2025 headcount disclosure to regulators.

Beyond that round of cuts, Meta closed another 6,000 open roles and is redirecting roughly 7,000 workers into newly created AI teams. Taken together, the moves touch close to 20% of everyone on the payroll.

Meta is not alone in leaning on AI to reshape headcount. JPMorgan Chase chief executive Jamie Dimon said this month that AI had already cut staffing by as much as 40% in some departments, a sign of how fast automated workforce decisions are spreading through corporate America.

Can a Company Legally Let AI Help Decide Who Gets Fired?

Federal law bars any employment process, automated or not, that produces a discriminatory result against a protected group without a valid business reason. California and New York City now also require bias testing before employers deploy automated tools in hiring, promotion or termination decisions, the specific gap the Meta plaintiffs say the company never closed.

The claims lean on a mix of federal and state law: the Family and Medical Leave Act (FMLA), the Americans with Disabilities Act (ADA), the Pregnancy Discrimination Act and the Pregnant Workers Fairness Act (PWFA). Together they bar employers from letting a facially neutral process, like a productivity score, produce a discriminatory outcome.

The scrutiny is reaching Meta on more than one front. Federal officials are separately pushing the company to join a government AI safety review, and a survey published this month found a majority of U.S. workers now back the idea of an AI wealth fund to offset the job losses the industry keeps tracing back to AI.

The Injunction Fight Ahead of July 22

Meta’s arbitration agreements route the underlying discrimination claims to private arbitration, and the plaintiffs accept that. Their request to the federal court is narrower: a preliminary injunction keeping them on payroll while an independent audit of the scoring process plays out and arbitration runs its course.

Once these terminations are finalized, the harm to Plaintiffs cannot be undone by money damages alone.

The complaint lists what a later court win could not undo: employer-subsidized health coverage lost mid-pregnancy or mid-treatment, leave rights that expire on a clock, unvested equity that disappears, and immigration status tied to employment. All 26 plaintiffs remain on Meta’s payroll for now, with separations scheduled to begin July 22.

Frequently Asked Questions

What Severance Are the Laid-Off Meta Workers Getting?

U.S. employees losing their jobs in Meta’s May cuts are receiving a base severance of 16 weeks’ pay, plus two additional weeks for every full year of service. Meta is also extending health insurance coverage for employees and their dependents for 18 months after separation.

Why Did the Plaintiffs Sue in Federal Court Instead of Going Straight to Arbitration?

Meta’s employment agreements require workers to arbitrate workplace disputes individually and privately. The plaintiffs argue that clause does not cover requests for emergency, temporary relief, so they filed in the U.S. District Court in Oakland to seek a preliminary injunction while their underlying discrimination claims proceed separately in arbitration.

What Is an Algorithmic Bias Audit Law?

It is a rule requiring employers to test automated tools used in hiring, promotion or termination decisions for discriminatory outcomes before deploying them. California and New York City have both adopted versions of this requirement recently, and the Meta lawsuit cites the company’s alleged failure to run that test as a separate legal violation.

When Do the Meta Layoffs Take Effect?

Meta began notifying roughly 8,000 employees in May, starting with staff in Singapore, who reportedly received layoff emails around 4 a.m. local time. Separations for the 26 plaintiffs, along with the rest of the affected workforce, are scheduled to take effect July 22.

Did Meta Confirm or Deny Using AI to Choose Who Was Laid Off?

Meta denies it. A company spokesperson told reporters, “Workforce management and organizational decisions were and are made by people, not AI.” The spokesperson called the lawsuit’s claims meritless and not based on facts.