26 Meta employees sue over AI-assisted layoff discrimination targeting workers on medical leave

Twenty-six anonymous Meta employees filed a lawsuit in Oakland federal court on July 14 alleging the company violated federal and state discrimination laws by using AI-assisted systems to target workers with disabilities, those on approved medical leave, and pregnant employees during the May 2026 mass layoff affecting nearly 8,000 staff. The complaint alleges Meta's "constellation of internal artificial-intelligence systems"—including Metamate (an LLM assistant), a "second brain" tracking worker communications, and productivity scores based on keystroke/screen scanning—created metrics that disadvantaged workers on protected absences.

The plaintiffs argue that metrics like AI token consumption and productivity scores could not account for workers on Family and Medical Leave Act (FMLA) protected leave or disability accommodations, creating a systematic bias against them. Meta's response: the company says workforce management decisions were made by people, not AI, and the claims lack merit. This defense mirrors Workday's stance in a related California case decided just last month, where a judge ruled the company must face claims of AI-powered discrimination in hiring.

The lawsuit underscores growing legal risk around AI-assisted personnel decisions. Plaintiffs are seeking a preliminary injunction to halt the remaining July 22 layoff notices pending independent audit and arbitration, plus claims under California and New York City laws requiring AI bias testing that Meta allegedly did not conduct.

For architects: this case signals that AI-adjacent workload in any talent-management system—even when the final cut decision is human-made—can face strict liability under anti-discrimination law if the inputs systematically penalize protected statuses. Bias by proxy (metrics that correlate with disability, pregnancy, or leave-taking) courts are treating as seriously as explicit protected-class logic.