Meta is now defending a lawsuit that goes straight at one of tech's touchiest questions: what happens when an algorithm helps decide who loses their job. Twenty-six current and former employees caught in the company's May layoffs allege that Meta's AI-driven selection process disproportionately picked workers who'd taken legally protected leave.
The allegations
The suit was filed this month in federal court in Northern California, and it stems from Meta's roughly 10% reduction in force earlier this year. The complaint says the termination list wasn't built "through the considered judgment of managers who knew the work." Instead, plaintiffs say, Meta leaned on "a constellation of internal artificial-intelligence systems" to score, rank, and select who got cut.
The individual stories are hard to read. A scientist says she was selected while on pre-birth pregnancy leave. A manager says he was demoted after medical leave, then chosen for layoff weeks into his second one. An engineer says his rating dropped because of the "broken time" an injury carved out of his working year.
Why the inputs matter
The legal core here is about data, not intent. The systems allegedly scored people on performance ratings, calibration scores, productivity metrics, "AI-native" ratings — even AI-token consumption, a measure of how heavily someone used the company's own AI tools.
See the trap? Someone out on protected medical or family leave can't accumulate those numbers. Neither can someone whose output shrinks because of a disability. And Meta, the plaintiffs allege, didn't "neutralize" the inputs to account for protected absences or pull leave-takers out of the selection pool. The scoring, per the lawsuit, "in effect penalized the employees for exercising their legal rights."
The claimed violations span most of the big federal protections: the ADA, the FMLA, the Pregnancy Discrimination Act, the Pregnant Workers Fairness Act, and Title VII. The workers want a preliminary injunction stopping Meta from finalizing their separations.
Meta's response
Meta rejects the premise flat out. The claims "lack merit and are not based on facts," a spokesperson said — "workforce management and organizational decisions were and are made by people, not AI."
So that's the factual fight: how much did automated scoring actually shape the layoff list, and how much human judgment sat on top of it?
The bigger picture
Whatever happens in this case, the timing is uncomfortable for the whole industry. Companies everywhere are quietly folding AI metrics into performance management — including, notably, measures of how much employees use AI itself. This lawsuit asks whether those systems can survive contact with employment law written decades ago to protect exactly the people most likely to fall through statistical cracks: the pregnant, the sick, the injured, the caregiving.
Discrimination statutes don't care whether the decision-maker was a manager or a model. If an input systematically disadvantages people who exercised a legal right, the tool built on top of it inherits the problem. Employment lawyers and HR teams will be watching discovery closely — because what it reveals about how Meta's list got made could set the template for every algorithmic layoff that follows.
Image: Nastuh Abootalebi, via Unsplash
