Twenty-six current and former Meta employees filed a federal lawsuit on July 15 alleging the company used what the complaint calls “a constellation of internal artificial-intelligence systems” to select workers for a roughly 10 percent reduction in force in May, and that the scoring disproportionately swept up employees who had taken or requested protected leave. The suit, filed in the U.S. District Court for the Northern District of California, names three representative plaintiffs: a scientist selected for layoff while on pre-birth pregnancy leave, a manager demoted then laid off during a second medical leave, and an engineer whose performance rating was lowered because of “broken time” tied to a work injury.
The case matters beyond Meta because it targets mechanics HR platforms have quietly normalized: performance ratings, calibration scores, productivity metrics, and even “AI-token consumption” feeding into layoff models, exactly the kind of automated scoring HR tech vendors have marketed as objective. The complaint alleges Meta never neutralized those inputs for employees who, by definition, could not accumulate normal output while on leave, effectively converting a protected legal status into a negative variable inside the model. Meta says workforce decisions “were and are made by people, not AI,” but the complaint argues the AI systems shaped who reached a human reviewer in the first place, and on what terms.
The original insight for HR leaders: the legal exposure here is not in having an algorithm assist a layoff decision, it is in feeding leave-affected metrics into that algorithm without a documented neutralization step. Any HCM system scoring workforce reductions on output-based inputs needs an audited exclusion for FMLA, ADA, and pregnancy-related leave built in before the next reduction in force, not retrofitted after a complaint is filed.
