Twenty-six anonymous ex-Meta employees sued the company in federal court in Oakland, claiming a stack of internal AI tools, not a manager’s judgment, decided who lost their job in May’s 8,000-person purge.
The Doe v. Meta complaint, filed in the Northern District of California, names a system called “Metamate,” employee-built “second-brain” agents, keystroke and activity monitors, and an AI-token-usage dashboard that graded staff as “AI Native,” “AI First,” or “AI Enabled.”
The math is the problem. Performance ratings, calibration scores, and AI-token consumption can’t accumulate for someone on FMLA leave or dealing with a disability that slows output. The plaintiffs say that structural gap, not merit, put them on the list, and it’s reportedly the first suit against a U.S. employer testing whether an algorithmic layoff pipeline violates the ADA and FMLA, the disability and family-leave protections every mid-size employer already has to honor.
For founders, this is the bill coming due for AI in HR ops. If you’ve plugged AI-usage telemetry into performance calibration, you need a documented human sign-off step and an audit trail showing a manager overrode the score for anyone on protected leave. That’s the cheap fix next to defending a class claim in arbitration.
Meta’s defense is flat denial: “Workforce management and organizational decisions were and are made by people, not AI.” That’s a testable claim, and discovery will pick it apart line by line. A federal judge now decides if it survives a motion to dismiss.
— Nathan Zakhary