A group of 26 former Meta employees has filed a federal lawsuit in Oakland, California, accusing the company of using AI systems to disproportionately target vulnerable workers, those with disabilities, serious medical conditions, or on approved medical and family leave, during its mass layoffs this year. The plaintiffs are asking the court to block Meta from completing the cuts, which are set to begin on July 22, while they pursue their claims in private arbitration, and Meta's answer so far is that people, not AI, made the final decisions about who was let go.
The context is one of the larger workforce reductions in recent tech history. Meta moved to cut roughly 10 percent of its global staff this year, close to 8,000 people, and according to the suit it relied on AI powered performance tracking to help decide who ended up on the list. The plaintiffs describe those systems as scoring workers on highly flawed metrics, leaning heavily on digital productivity signals, the kind of measures that reward whoever is most active at the keyboard rather than whoever is doing the most valuable work.
The complaint names specific internal tools, which is part of what makes it striking. It points to Metamate, an AI assistant, to what it calls an employee trained second brain that tracked workers' communications and documents, and to a productivity score assembled by scanning keystrokes, screen content, emails, and browser history. Whether or not those descriptions hold up, they sketch a picture of pervasive workplace monitoring feeding an automated ranking, and the plaintiffs argue that ranking is exactly where the discrimination entered.
The legal heart of the case is disparate impact. The argument is not that anyone typed a rule saying penalize sick employees, but that scoring people on raw digital activity will systematically flag workers who were less active for reasons the law protects, because they were on medical leave, managing a disability, or dealing with a serious illness. A health circumstance becomes a low productivity score, and a low score becomes a layoff. Meta's defense, that humans made the final calls, runs straight into the central question of AI in the workplace, whether a system that ranks and effectively recommends who should be cut carries responsibility when a person merely signs off on what it produced.
Why it matters reaches well beyond Meta. This is one of the first high-profile tests of AI making, or heavily shaping, high-stakes employment decisions, and it sits at the intersection of two things people are already uneasy about, workplace surveillance and algorithmic bias. However it resolves, and these are allegations headed to arbitration rather than established facts, it will start to define how much a company can lean on AI to decide who keeps a job and how much accountability it holds when the outcomes fall hardest on protected groups. Every company now wiring AI into hiring and firing is about to learn that those decisions come with a courtroom attached.