Meta maintains that its workforce management decisions are 'made by people'
Former and current employees at Meta have filed a lawsuit against the tech giant alleging its mass layoff in May 2026 disproportionately targeted workers on protected medical and family leave, claiming that an artificial intelligence system, not human managers, made the selections.
The complaint, filed in the US District Court for the Northern District of California, names Meta as defendant and accuses it of using internal AI tools to score, rank, and select employees for termination without adjusting those tools to account for time spent on legally protected leave.
Meta announced in April 2026 that approximately 8,000 employees, or roughly 10% of its workforce, would be laid off as the company sought to remake itself into what it called an "AI-first" organisation.
The 26 plaintiffs, proceeding anonymously, allege the selection process relied on inputs including AI-token consumption, keystroke and activity monitoring data, algorithmic performance rankings, and scores from an internal tool called "Metamate."
"Meta did not neutralise those inputs for protected leave; did not exclude protected-leave-takers or accommodation-seekers from the selection cohort; and did not pause the system for the individualised, leave- and accommodation-neutral review that the law requires," the complaint states.
"The result was that employees who took protected leaves were disproportionately selected for layoff, based on scoring that not only failed to account for their protected leaves, but in effect penalised the employees for exercising their legal rights to these leaves."
One scientist was notified her role had been eliminated while she was on approved pre-birth maternity leave. She was selected for redundancy a day before her water broke, and just two days before she gave birth.
Another employee's own performance review, the complaint alleges, documented that his demotion "followed his return from medical leave."
A third employee's manager tied his lowered rating directly to what she described as the "broken time" when an injury had prevented him from working.
Meta's monitoring program
The complaint further alleges that Meta deployed a workforce-wide monitoring program – later announced internally as the "Model Capability Initiative" – to capture keystrokes, screen content, browser history, messaging, email, and in some cases voice, video, and location data from company-issued devices.
More than 1,000 employees reportedly signed an internal petition against the program. The complaint notes that on June 22, it was publicly reported that Meta had paused the initiative after an internal security incident exposed employees' private conversations and performance data across the entire company.
Plaintiffs further allege that some employees were actively discouraged from taking leave. One plaintiff's manager allegedly warned him that submitting a leave application "will definitely" result in termination.
The manager then proposed what he called a "win-win," offering a favourable performance letter in exchange for dropping the leave, and refused to put the threat in writing to "avoid a paper trail."
The 21-count complaint invokes the Family and Medical Leave Act, Title VII of the Civil Rights Act, the Americans with Disabilities Act, the Pregnant Workers Fairness Act, and a range of state laws across California, Washington, New York, Illinois, Pennsylvania, Florida, and the District of Columbia.
It also cites California's newly effective regulations on automated-decision systems, which took effect on 1 October 2025.
Review of AI selection process sought
Because Meta's employment agreements require individual arbitration, plaintiffs are not seeking class-wide relief.
Instead, they are asking the court to issue a preliminary injunction halting their terminations pending an independent audit of the AI selection process, with individual claims to proceed in arbitration.
Meta has denied the allegations. A spokesperson told CNBC that the "claims lack merit and are not based on facts."
"Workforce management and organisational decisions were and are made by people, not AI," the spokesperson told the news outlet.
