Meta sued by 26 workers over AI-driven layoff selection

Something worth pausing on in The Guardian's reporting: 26 Meta employees have filed suit in federal court in Oakland, alleging the company used internal AI systems to flag them for the roughly 8,000-person workforce cut Meta announced in May. The plaintiffs are all still on payroll, with separations set to begin July 22, which is part of what gives the case its unusual shape.

The specifics are what make this more than the usual algorithmic-bias abstraction. The complaint names "keystroke and activity-monitoring data, AI token-usage dashboards and algorithmically assisted performance rankings" as the tooling that produced the scores, and argues those scores "by design, cannot be accumulated by an employee who is on protected medical or family leave, or whose output is reduced by a disability." The plaintiff group, as CBS News summarized, includes eight women on maternity or pregnancy leave, four men on parental leave, and one woman on family bereavement leave. One plaintiff, a scientist on approved pre-birth pregnancy leave, was reportedly notified of her layoff two days before she gave birth; a manager on medical leave was let go 16 days into his time off.

Why this matters beyond the 26 named workers: the plaintiffs are invoking Title VII disparate-impact doctrine, the Family and Medical Leave Act, the Americans with Disabilities Act, the Pregnancy Discrimination Act, and the Pregnant Workers Fairness Act, and they explicitly claim Meta failed to test its AI systems for bias in violation of recently adopted California and New York City laws. That is the ruleset under which every large employer piloting algorithmic performance tooling now has to defend its process. If a court entertains disparate-impact theory applied to algorithmic layoff scoring, HR-tech buying decisions across the industry get more expensive very quickly.

The honest caveat is that this is a filed complaint, not a finding, and Meta's position is that the claims "lack merit" and that "workforce management and organizational decisions were and are made by people, not AI." What the reporting does not give you is which Meta systems by product name produced which scores, or what human review sat between the algorithm and the final termination list. Discovery is where that gets settled.

The forward-looking piece is that vendors selling audited, bias-tested performance AI, and the employment lawyers now templating this kind of case, both have a lot more work in front of them whichever way this one lands.