# Meta Used AI to Rank Workers for Layoffs — And It Flagged Everyone on Leave > Source: > Published: 2026-07-15 13:06:53+00:00 # Meta Used AI to Rank Workers for Layoffs — And It Flagged Everyone on Leave Twenty-six anonymous Meta employees filed suit in Oakland federal court alleging the company's AI-powered ranking systems flagged workers on medical, pregnancy, or parental leave for termination during May's 8,000-person layoff. The suit cites the ADA, FMLA, Pregnancy Discrimination Act, and California and NYC AI-bias laws. Meta says 'decisions were made by people, not AI.' Twenty-six anonymous Meta employees filed suit in Oakland federal court on July 13, alleging the company's AI-powered ranking systems used productivity and [token](/glossary/token)-usage metrics that disproportionately flagged workers on medical, pregnancy, or parental leave during May's 8,000-person layoff. The plaintiffs include a data scientist selected for termination two days before giving birth and a manager on approved pregnancy disability leave. The suit cites the Americans with Disabilities Act, the Family and Medical Leave Act, the Pregnancy Discrimination Act, and both California and New York City AI-[bias](/glossary/bias) laws. Meta's response: "workforce management decisions were made by people, not AI." That defense — "humans made the final call" — has become the standard corporate response to AI bias claims. It rarely holds up in court, and this case is shaping up to test exactly where the line is. ## How the System Worked According to the complaint, Meta's internal ranking engine pulled productivity signals — code commits, diff volume, token usage in AI-assisted workflows, meeting attendance — and generated a ranked list for each team. Managers received the list and made "final" selection decisions within tight percentage quotas. The problem: productivity metrics don't account for leave. If you're on FMLA for six weeks, your commit count is zero. If your AI token usage drops because you're on pregnancy disability, the algorithm doesn't know why. It just sees the drop. The plaintiffs argue the system created an illegal proxy for protected status. You weren't flagged because you were pregnant. You were flagged because you had zero commits for two months — because you were pregnant. The distinction matters legally, and the ADA and FMLA don't care about the AI's intent. They care about the effect. ## Why This Case Matters This isn't the first AI-bias employment suit, but it's the highest-profile one to date. Meta's 8,000-person layoff gives the plaintiffs a large sample size. The 26 named plaintiffs represent a broader class that could number in the hundreds. And the combination of federal and state AI-bias laws — California's and NYC's in particular — gives the plaintiffs multiple legal hooks. If the case survives summary judgment and reaches discovery, Meta's internal ranking algorithms become evidence. That's a precedent no big tech company wants to set. But the bigger question is whether "humans made the final call" works as a defense. If the AI ranked you at the bottom and the manager had to cut 15% of the team, does it matter that a human technically signed off? The plaintiffs don't think so. The courts are about to weigh in. #### Q: What is Meta accused of doing? #### Q: How many people were laid off? #### Q: What laws does the suit cite? ## Related Articles Get AI news in your inbox Daily digest of what matters in AI.