A lawsuit filed in California federal court in July challenges Meta's May decision to lay off 8,000 employees—roughly 10 per cent of its global workforce—on grounds that the company deployed artificial intelligence systems to identify candidates for termination in ways that disproportionately harmed workers exercising their legal right to take medical, parental or family leave. Twenty-six anonymous Meta staff members named in the complaint allege that the technology company's selection methodology was fundamentally flawed because it failed to accommodate or account for protected absences, thereby creating a process that systematically disadvantaged a specific class of workers.
The mechanics of Meta's layoff system, according to the lawsuit, relied on multiple algorithmic tools and data-collection mechanisms to rank and score employees. These included keystroke monitoring, activity-tracking dashboards, AI token-usage metrics, and algorithmically-assisted performance rankings. The core allegation is that these measurement systems operated continuously without pause or adjustment for employees on protected leave. Because workers taking medical, parental or family leave necessarily reduce their measured output during their absence—a temporary circumstance mandated by law—the algorithms recorded this absence as diminished performance. For employees whose performance was flagged as substandard by these systems, the risk of layoff selection increased significantly.
The 26 plaintiffs each took some form of legally protected leave and sought or received workplace accommodations for disabilities. Eight were women who had taken pregnancy-related or maternity leave; four were men on parental leave; one was a woman who took leave to care for a sick family member and later for bereavement. All have been formally notified of their termination, though separations had not been finalised at the time the lawsuit was filed in mid-July, with final departures scheduled to begin a week later. This timing allowed the plaintiffs to seek an emergency injunction to preserve their employment status pending arbitration or settlement.
The lawsuit invokes several federal and state protections that Meta allegedly violated. The Family and Medical Leave Act guarantees eligible workers up to 12 weeks of unpaid, job-protected leave annually. The Americans with Disabilities Act requires employers to provide reasonable accommodations to workers with disabilities unless doing so creates undue hardship. The Pregnancy Discrimination Act explicitly forbids terminating or demoting workers based on pregnancy, childbirth or related conditions. The Pregnant Workers Fairness Act, a more recent statute, mandates that employers provide reasonable accommodations to pregnant employees. The complaint contends that Meta's AI-driven process circumvented the protective spirit of all these laws by treating temporary, legally-mandated leave as evidence of poor performance.
One particularly striking allegation concerns a male employee with a diagnosed serious health condition and disability approved by Meta's own healthcare provider. According to the lawsuit, his manager actively discouraged him from taking the leave to which he was legally entitled, warning explicitly that doing so would trigger his selection for layoffs. He was not offered any accommodation for his disability, the complaint states. This account, if proven, would suggest that Meta's human managers understood and warned about the adverse effect the AI system would have on workers on leave, raising questions about whether company leadership knowingly allowed a flawed process to proceed.
Meta has rejected the allegations, issuing a statement asserting that the claims "lack merit and are not based on facts," and insisting that "workforce management and organisational decisions were and are made by people, not AI." This defence—that humans made the final decisions—mirrors arguments made by other technology companies facing AI discrimination lawsuits. However, the plaintiffs' legal team counters that the question is not whether humans made decisions, but whether those decisions relied on algorithmic scores that were themselves tainted by the failure to account for protected leave.
The lawsuit relies in part on the legal doctrine of disparate impact, a civil rights principle holding that employment practices that appear neutral on their face can still be unlawful if they disproportionately harm workers in a protected class and are not necessary for job performance. The plaintiffs argue that Meta's algorithm-assisted selection process created disparate impact by design: because women disproportionately take pregnancy and caregiving leave, a system that penalises reduced output during leave falls more heavily on female workers than male workers. The doctrine traces its roots to a landmark 1971 Supreme Court decision and is codified in Title VII of the Civil Rights Act.
This legal theory carries particular significance in the current political environment. The Trump administration has moved to curtail disparate impact enforcement, instructing federal agencies to deprioritise such claims and arguing that the doctrine undermines "meritocracy" and wrongly assumes that workforce imbalances automatically signal discrimination. The administration has directed the Equal Employment Opportunity Commission to drop some pending cases based on disparate impact theory. These policy shifts have led some observers to question whether disparate impact liability remains a viable legal tool.
However, the Meta lawsuit demonstrates that workers retain important protections despite federal policy shifts. While the EEOC may reduce its enforcement efforts, private lawsuits brought by workers themselves remain permissible, and several states have enacted their own laws explicitly prohibiting disparate impact discrimination. The plaintiffs' legal team is not relying solely on federal enforcement agencies; they are asserting workers' independent right to sue. This distinction matters for companies operating in multiple jurisdictions, including those with robust state-level employment protections.
The plaintiffs are seeking an emergency injunction to maintain their employment status pending arbitration or resolution, rather than pursuing immediate financial damages. Their lawyers emphasise that allowing the layoffs to proceed would cause irreversible harms: loss of employer-sponsored health insurance at a time when several plaintiffs are pregnant, postpartum, or undergoing medical treatment; forfeiture of time-limited leave entitlements that cannot be recovered; loss of unvested equity compensation; and potentially serious immigration consequences for any non-citizen workers affected.
For technology companies and other employers globally, including those in Southeast Asia, the lawsuit presents a cautionary case study in the intersection of algorithmic decision-making and employment law. As firms increasingly adopt AI systems for workforce management—from hiring and performance evaluation to retention and reduction decisions—they face mounting legal exposure if those systems are not carefully designed to incorporate legal requirements and protected statuses. Meta's public statement that decisions were "made by people, not AI" suggests the company believes this framing provides legal cover, but the plaintiffs' case asserts that reliance on a flawed algorithmic foundation for those human decisions does not cure the underlying violation.
The case also raises questions about corporate transparency and worker agency in the age of algorithmic management. The plaintiffs describe systems that continuously monitored keystrokes and activity, generated opaque performance scores, and informed consequential decisions about workers' livelihoods, yet many affected workers apparently did not fully understand how they were being evaluated or why they were selected for layoffs. This opacity—common in many algorithmic systems—makes it difficult for workers to contest decisions or advocate for themselves. Malaysian and Southeast Asian workers, many of whom lack strong collective bargaining arrangements or union representation in the tech sector, may face similar challenges if international technology companies import such systems into regional offices without meaningful safeguards.
