In five days, Meta was set to quietly finalize another layoff round. Instead, the company is now facing a federal lawsuit arguing that an algorithm, not just management, decided who stayed and who left.
Twenty-six current and former employees, filing anonymously as Does 1 through 26, sued Meta in the U.S. District Court for the Northern District of California. They claim that when Meta cut roughly 8,000 jobs last spring, it leaned on internal AI tools to rank staff — including the Metamate assistant, a performance system called Checkpoint that rewarded heavy AI usage, plus keystroke logs, browser activity, and per-employee AI token consumption.
The trouble, the suit says, is that none of those metrics adjusted for people who were on medical leave, parental leave, or caring for a family member at the time the scores were calculated. Less logged activity meant a lower score, and a lower score meant a higher chance of being cut, regardless of whether the absence was legally protected.
The plaintiffs cite the ADA, the Pregnancy Discrimination Act, and the Family and Medical Leave Act, along with newer California and New York City rules requiring bias testing of algorithmic hiring and employment tools. They aren't asking for damages yet, just an emergency injunction pausing the terminations scheduled for July 22 until an independent audit of the system can be completed. Individual claims will move to private arbitration under their employment contracts.
Meta calls the allegations baseless and says the final calls were made by people, not software. Still, lawyers tracking the case describe it as the first major U.S. lawsuit to challenge layoffs specifically as a product of AI decision-making, and a test of whether discrimination law still applies once a company lets an algorithm do the sorting.
