A group of former Meta employees who claim they were included in a round of layoffs because an artificial intelligence tool and didn't factor their disability accommodations or protected leave into a productivity analysis cannot qualify for an injunction reinstating them to their jobs, Meta argued in California federal court.

In a brief filed Monday, Meta Platforms Inc. said the 25 ex-employees haven't improved their legal argument since they lost their emergency application for a temporary restraining order last month. U.S. District Judge William H. Orrick's denial of the application "is significant, and should inform the analysis here," Meta argued.

"Plaintiffs now repackage substantially the same arguments and evidentiary record, supplemented with expanded legal theories that do nothing to cure the fundamental deficiencies this court has identified," Meta said. "The answer remains the same: Plaintiffs' claims belong in arbitration, their alleged harms are compensable through monetary relief, and this court should not grant the extraordinary remedy of a preliminary injunction."

Judge Orrick denied the ex-workers' TRO bid July 17, saying the workers didn't prove, at that stage of the litigation, that they'd suffer irreparable harm if they weren't immediately reinstated to their jobs. Loss of employment and healthcare is not an irreparable injury, because it can be remedied "through damages, back pay, or other comparable relief in the arbitration process," the judge said.

The strongest argument for irreparable harm came from four ex-employees who had their work visas sponsored by Meta, the judge said. He said that at the preliminary injunction stage of the litigation, he would examine the expanded case record to decide whether those workers deserved immediate reinstatement. He ordered Meta to submit a declaration answering whether it used AI when selecting those four workers for inclusion in its May reduction in force.

Meta submitted that declaration July 23 via human resources executive Linh Doan, claiming the company didn't use AI at all when preparing the RIF. "Selection decisions were made by human business leaders," Doan said.

"AI was not used to conduct the RIF. AI did not select anyone for termination. AI did not assist anyone in making selection decisions," Doan said. "AI was not used to calculate performance scores, to rank employees, or to generate, influence, or recommend selection criteria."

The company doubled down on that assertion Monday, saying "human business leaders" based their decisions on whom to lay off on "neutral criteria" such as workers' tenure and skills.

"These leaders were only authorized to make selection decisions by applying the specific, pre-determined permissible criteria," Meta said. "Those permissible criteria did not include leave status, leave history, disability status, or accommodation requests, or any other protected characteristic."

The ex-workers sued Meta July 13, claiming they were included in the May RIF because the company didn't factor their disability accommodations or leave into its AI-assisted pre-layoff analysis of their productivity. They accused the company of violating state and federal laws banning disability discrimination in the workplace and are seeking reinstatement and damages. Meta has denied the allegations and argued that the workers should arbitrate their claims.

Representatives of the parties were not immediately available for comment Tuesday.

The former employees are represented by Andrea Mazingo of Lumen Law Center PC, by Barbara E. Cowan of Workplace Advocates and by Rebecca Houlding of Houlding Law PC.

Meta is represented by Erin M. Connell, Lauren M. Goldsmith and Dixie Tauber of Orrick Herrington & Sutcliffe LLP.

The case is Does 1 et al. v. Meta Platforms Inc., case number 3:26-cv-07122, in the U.S. District Court for the Northern District of California.

--Editing by Lief Nielsen.

For a reprint of this article, please contact reprints@law360.com.