The issue focuses on layoffs in May 2026, when Meta cut about 8,000 employees, or about 10% of its global workforce, as it repositions itself as an “AI-first” organization. The plaintiffs, whose identities are redacted from court filings, include engineers, managers, researchers, designers, and at least one director with direct access to the AI platform in question.
They claim that Meta’s selection process relied on an internal large-scale language model called Metamate that was trained on employee communications and documents, and relied on algorithmic productivity scores derived from keystroke activity, browser history, email data, internal records of AI token consumption and AI-assisted performance review tools. Importantly, the complaint alleges that this scoring framework does not take into account time spent on legally protected leave. This meant that employees who took maternity, paternity, or medical leave were effectively penalized for less production during the period they were legally entitled to take leave.
Nearly half of the plaintiffs say they were fired immediately after being placed on protected leave. “Meta did not base its list of layoffs on the prudent judgment of knowledgeable management,” the complaint states, placing the case squarely at the intersection of employment discrimination law and algorithmic accountability.
