A federal judge has refused to put the brakes on Meta Platforms’ planned layoffs, even as 26 workers claim the tech company used artificial intelligence tools to unfairly single them out for job cuts because of disabilities or medical leave.
U.S. District Judge William Orrick, based in Oakland, California, issued a written ruling Friday stating he would not prevent Meta from moving forward with the layoffs starting July 22, while the employees’ groundbreaking legal claims are resolved through private arbitration.
In his decision, Orrick determined that the workers failed to demonstrate they would suffer what the law calls “irreparable harm” — the legal standard required before a judge can issue an emergency order to stop the layoffs.
Neither Meta nor attorneys representing the employees responded to requests for comment before deadline. Meta has denied any wrongdoing and maintained that human beings — not AI systems — made the final decisions about who would be laid off.
Back in May, Meta announced that nearly 8,000 employees — roughly 10% of its worldwide workforce — would be losing their positions as the company continues to pour resources into artificial intelligence development.
The lawsuit, filed earlier this week, alleges that Meta used AI-driven tools to measure worker productivity and AI usage, which put employees at a disadvantage if they had been away from work due to medical conditions or family caregiving responsibilities. Workers also say the company factored in performance reviews that partly graded employees on how much they embraced AI technology.
Legal experts note this case appears to be the first of its kind in the United States — a direct legal challenge to a major corporation’s alleged use of AI in carrying out mass layoffs.
The employees had asked Judge Orrick for a temporary restraining order to pause the layoffs while their claims go through arbitration. A request for a longer-term preliminary injunction is still pending, and Orrick indicated during a Thursday hearing that he expects to rule on that matter next month.
During Thursday’s hearing, attorneys for the workers argued that beyond their paychecks, the employees stood to lose valuable stock options and health insurance coverage — putting at risk medical care for pregnancies and other ongoing conditions.
“There’s no do-over for bonding with a new baby or giving birth or having active medical treatment,” said attorney Barbara Cowan, addressing the judge directly.
Meta’s attorney, Erin Connell, pushed back, arguing that workers would only be losing employer-subsidized insurance — not their health coverage entirely. Connell said those types of financial losses could be recovered later if the employees prevail in arbitration.
The workers contend that while Meta’s employment agreements require disputes to go through individual arbitration, those agreements still allow employees to seek temporary court relief — an argument at the heart of this case.
Arbitration agreements are standard practice at most large companies, requiring workers to handle workplace grievances individually rather than joining class-action lawsuits. Supporters say arbitration is faster and less expensive than going to court, while opponents argue it tends to benefit employers and makes it harder for workers to pursue legitimate claims.
The plaintiffs — who filed the case anonymously — include engineers, managers, researchers, and designers. They were notified of their layoffs in May. While many are scheduled to be officially terminated on July 22, others won’t lose their jobs until later in July or August, according to court documents.
Meta noted in court filings that although the affected workers are still technically on the payroll, they lost access to company systems on May 20 and have not performed any work since then.
According to the lawsuit, Meta used several internal AI-assisted systems to evaluate and rank employees for termination. These included a large language model assistant called “Metamate,” an employee-built “second brain” that monitored workers’ communications and documents, and a productivity scoring system that analyzed keystrokes, screen activity, emails, and browsing history.
The plaintiffs claim Meta did not pause these tracking systems during vacations or legally protected leave, causing employees’ AI adoption scores — which were used as part of the layoff selection process — to drop during those absences.
