Twenty-six current and former Meta employees are suing the company in federal court, alleging that AI-assisted systems played a central role in deciding who got laid off. The lawsuit, filed in July 2026 in Oakland, California, claims that tools including productivity scoring, AI token usage tracking, and performance ranking algorithms were used to select employees for termination, and that these systems disproportionately targeted workers on medical, parental, or family leave.
What happened at Meta
The lawsuit lands against the backdrop of a significant workforce reduction at Meta Platforms. In May 2026, roughly 8,000 employees, about 10% of the company’s total headcount, were notified that their jobs were being eliminated. The scheduled termination date for affected workers was set for July 22, 2026.
The plaintiffs moved fast, seeking a temporary restraining order to block the layoffs before that date. US District Judge William Orrick denied the request, but his reasoning was telling. He acknowledged the fundamental complexity of proving AI’s influence in employment decisions, noting that the plaintiffs “were not in the rooms where it happened.”
Among the AI tools named in the complaint is “Metamate,” described as a large language model assistant, alongside various productivity monitoring technologies. The lawsuit alleges these systems violated bias-testing requirements under both California state law and New York City regulations.
A hearing on a preliminary injunction is scheduled for August 24, 2026. That proceeding could establish early precedent for how courts evaluate claims of algorithmic discrimination in the workplace.
What this means for tech companies and investors
The suit invokes bias-testing requirements under California and New York City laws, two jurisdictions that have been at the forefront of regulating automated decision-making in employment. New York City’s Local Law 144, which requires bias audits for automated employment decision tools, has been on the books since 2023 but enforcement has been uneven.