Meta faces lawsuit over alleged discriminatory AI in 8,000-person layoff

Former and current Meta employees filed a federal lawsuit in California alleging that the company relied on internal AI systems to generate employee layoff lists during a May reduction of 8,000 workers. According to the complaint, the AI systems targeted employees with disabilities and those on protected medical, family, or parental leave at disproportionate rates. The selection criteria the systems allegedly applied included performance ratings, productivity, work output, and measured AI usage. One plaintiff was notified of termination just two days before giving birth, the complaint states. Meta responded through a spokesperson, denying the allegations and asserting that humans make all personnel decisions at the company. The plaintiffs are seeking a preliminary injunction to preserve their employment status while arbitration proceeds. The case raises questions about the role of AI in automating high-stakes workforce decisions and whether algorithmic selection can discriminate against protected classes even when not intentionally designed to do so.

The lawsuit brings into sharp focus a growing tension in AI deployment: while machine learning systems are often justified as objective and impartial, they can encode or amplify existing biases in training data or selection logic. Meta's alleged use of AI to generate layoff lists—reportedly without adequate human review at the decision-making stage—sits at the intersection of employment law and algorithmic fairness. The complaint's specific claim that workers on protected medical, family, or parental leave were targeted at disproportionate rates suggests the AI system either directly penalized absence or correlated it with other metrics in ways that produced discriminatory outcomes. Meta's denial that humans make all personnel decisions conflicts with the plaintiffs' core allegation, making the factual record in discovery—what the AI system actually flagged, how humans acted on those flags, and whether override mechanisms existed—central to the case's outcome.