Meta sued for allegedly using AI to target employees on protected leave in layoffs
Meta employees sued the company alleging that AI systems used in May 2026 layoffs disproportionately targeted workers on protected medical or family leave, including one employee told of her termination two days before childbirth. A federal judge rejected a request to halt the layoffs, finding insufficient evidence Meta actually used AI in termination decisions.
What happened
On May 20, 2026, Meta laid off approximately 8,000 employees, or 10% of its workforce. According to a lawsuit filed by 26 current and former employees in July 2026, Meta used a 'constellation of internal artificial-intelligence systems' including its proprietary AI assistant Metamate to create a termination list.
The complaint alleges that the AI systems relied on metrics such as productivity, AI-token use, performance ratings, calibration scores, and 'AI-native' ratings to determine who would be laid off. The plaintiffs claim these metrics systematically disadvantaged employees on protected medical or family leave, or whose output was reduced by a disability, because such employees could not accumulate these metrics while on leave.
Reported impact
The lawsuit names 26 current and former Meta employees who were notified of their layoffs. According to the complaint, one plaintiff was on preapproved pregnancy leave and was told she was being laid off two days before giving birth. Another plaintiff was on medical leave for a 'serious health condition'.
The plaintiffs allege they were disproportionately selected for layoff because they exercised their legal rights to protected leave, violating discrimination laws. The plaintiffs sought a preliminary injunction to maintain their employment status pending arbitration and to prevent Meta from completing their separations.
Index analysis
The case raises questions about how AI systems can perpetuate discrimination in employment decisions. The lawsuit alleges that by design, the AI metrics used could not be accumulated by employees on protected leave, effectively penalizing them for exercising their legal rights.
A federal judge rejected the plaintiffs' motion for a temporary restraining order on July 17, 2026, finding insufficient evidence that Meta actually used AI in termination decisions, noting that plaintiffs 'were not in the rooms where it happened.' Meta's director of human resources business partner enablement declared that termination decisions were 'made by human business leaders based on documented, neutral criteria.'
Governance controls
Meta stated that workforce-management decisions were made by people, not AI, and that the allegations 'lack merit and are not based on facts.' Meta submitted a declaration from its human resources director denying that AI was used in the termination process.
Judge William H. Orrick ordered Meta to provide a declaration explaining the termination of four visa-sponsored plaintiffs by July 23, 2026, and noted that discovery during arbitration would be necessary to test the credibility of Meta's declarations.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.