Meta is facing a significant legal challenge from 26 former employees who contend the company deployed artificial intelligence tools to identify and remove staff members on protected leave when it eliminated approximately 8,000 positions — roughly one-tenth of its global workforce — in May. The lawsuit, filed in federal court in Oakland, California in mid-July, paints a troubling picture of algorithmic decision-making that allegedly bypassed fundamental employment protections while the company presented the cuts as merit-based personnel management.
The case hinges on the mechanics of how Meta's systems evaluated workers. According to the complaint, the company relied on keystroke monitoring, activity-tracking data, AI token-usage metrics, and algorithmically-assisted performance ratings to determine which employees would be terminated. Critically, the employees argue, these measurement systems were fundamentally incompatible with periods when staff members were absent for legitimate, legally protected reasons. Someone on maternity leave, medical leave, or taking time off to care for a sick family member would necessarily record lower activity metrics and productivity scores — not because of underperformance, but because they were not working during their protected absence.
The lawsuit presents a nuanced understanding of how ostensibly neutral systems can encode discrimination. The 26 plaintiffs — all of whom remain employed by Meta pending their scheduled separation on July 22 — collectively represent a pattern of vulnerability that affected those with particular caregiving responsibilities. Eight women had taken maternity or pregnancy-related leave; four men had taken parental leave; one woman had taken family care and bereavement leave. These demographic patterns matter legally because they suggest the algorithm's impact fell disproportionately on women, who statistically undertake more caregiving responsibilities than men.
One plaintiff's experience illustrates the alleged coercion embedded within the process. According to the lawsuit, this individual disclosed a serious health condition and disability approved by Meta's own medical provider. Rather than receiving accommodation or reassurance about job security, the employee was actively discouraged from taking medical leave by a manager who warned that doing so would trigger selection for termination. The employee ultimately did not take the leave, fearing the exact outcome that materialised — a decision that underscores how the threat of algorithmic retaliation can chill the exercise of legal rights.
Meta has rejected the allegations, stating in a brief response that the claims lack factual foundation and asserting that workforce decisions were made by people, not machines. This argument, however, misses the crux of the complaint: the question is not whether humans made final decisions, but whether they relied on data generated by systems that systematically disadvantaged protected workers. The lawsuit does not claim the algorithm achieved sentience and independently terminated staff, but rather that Meta used algorithmic outputs as the basis for human decision-making in a manner that violated federal and state employment law.
The legal framework supporting the case draws on multiple statutes. The plaintiffs invoke the Family and Medical Leave Act, the Americans with Disabilities Act, the Pregnancy Discrimination Act, and the Pregnant Workers Fairness Act — a relatively new statute that requires employers to provide reasonable accommodations to pregnant workers unless doing so creates undue hardship. These laws form a comprehensive safety net that Meta allegedly circumvented by designing systems that made it mathematically impossible for workers on protected leave to maintain competitive performance metrics.
The lawsuit also invokes the doctrine of disparate impact — a civil rights concept that holds employers liable for employment practices that are facially neutral but have a disproportionate adverse effect on protected groups. This legal theory carries heightened significance in the current political climate. The Trump administration has moved to deprioritise disparate impact enforcement across federal agencies, arguing that the doctrine undermines meritocracy and promotes unfounded assumptions about discrimination. The administration's order has influenced the Equal Employment Opportunity Commission's approach to case selection, with the agency dropping some discrimination complaints based on disparate impact theory.
However, the Meta case demonstrates that companies remain exposed to disparate impact liability regardless of the administration's enforcement priorities. Workers retain the right to pursue such claims independently if federal agencies decline to pursue them, and several states have incorporated disparate impact protections into their own employment laws. California, where the lawsuit was filed, has particularly robust employment discrimination statutes. The plaintiffs' legal team has argued that Meta's process of recording leave periods as reduced performance falls more heavily on women, whose greater propensity to take pregnancy and caregiving leave makes them more vulnerable to algorithmic selection.
The practical stakes of the lawsuit extend beyond the immediate question of employment status. The plaintiffs' legal team is seeking to preserve the current employment relationship pending arbitration, arguing that finalised terminations create irreversible harms. Once layoffs are complete, workers lose employer-subsidised health insurance at moments when coverage may be medically critical — during pregnancy, postpartum recovery, or active treatment for serious health conditions. Time-limited leave entitlements expire, unvested equity is forfeited, and immigration consequences may be triggered for visa-dependent employees, making reinstatement mathematically insufficient to remedy the injury.
The case raises broader questions about how artificial intelligence systems are deployed in workforce management across the technology industry and beyond. As companies increasingly turn to algorithmic tools to optimise personnel decisions, they face a fundamental design challenge: how to incorporate the complexity of employee protections into systems built for efficiency and quantification. The Meta lawsuit suggests that treating leave as simply another variable to be measured and ranked — rather than as a legally protected status requiring separate, individualised review — may constitute unlawful discrimination regardless of the sophistication of the underlying technology.
For Malaysian and Southeast Asian readers, the case offers instructive precedent as regional companies increasingly adopt AI-driven human resources practices. Employment protections for workers on medical, parental, and family leave exist across the region, with variations in scope and enforcement. The Meta litigation demonstrates that companies cannot offload responsibility for legal compliance to algorithms, nor can they use the opacity of AI systems as a shield against employment discrimination claims. As technology diffuses throughout the region's business landscape, these lessons about algorithmic accountability and worker protection will likely become increasingly relevant to companies operating across multiple jurisdictions.
