A US federal judge has allowed Meta Platforms to proceed with terminating 26 employees who contend that the company's artificial intelligence systems systematically targeted them for dismissal due to disabilities or approved medical absences. The decision by U.S. District Judge William Orrick of the Oakland, California federal court on Friday removes an immediate obstacle to layoffs scheduled to begin on July 22, though the judge's written reasoning leaves open the possibility of future reversals should additional evidence emerge regarding how AI shaped the termination decisions.
The workers had sought an emergency restraining order to freeze the dismissals while their claims proceed through private arbitration, arguing they faced immediate and irreparable harm by losing not only salaries but also stock options and employer health coverage tied to ongoing medical treatments. Judge Orrick found, however, that the plaintiffs had not demonstrated the exceptional legal threshold of irreparable injury necessary to justify such an extraordinary intervention before trial. His order signals skepticism about their ability to recover what they would lose through job termination, viewing the financial and benefits consequences as damages that could theoretically be remedied through monetary awards later if the workers ultimately prevail in arbitration.
The lawsuit filed this week represents what appears to be the first significant legal challenge in the United States against a major technology corporation for allegedly deploying artificial intelligence to identify and eliminate employees for workforce reduction. The case touches on accelerating tensions in Silicon Valley over the transparency and fairness of algorithmic decision-making in human resources, particularly as companies lean increasingly on AI to justify personnel cuts. The 26 anonymous plaintiffs—engineers, researchers, managers, and designers—allege that Meta relied on multiple interconnected AI systems to score, rank, and select workers for termination, creating what they characterise as a systematic mechanism for discrimination masked by technological opacity.
According to the complaint, Meta employed several AI-enabled tools in the selection process. The company used "Metamate," an internally developed large language model designed to function as an employee assistant, which tracked workers' communications and documents. The company also deployed a productivity-scoring system that analysed keystroke activity, screen content, email patterns, and browsing history to generate numerical rankings used in termination decisions. Crucially, the plaintiffs argue, these systems continued functioning and penalising workers during periods of approved vacation, medical leave, and other legally protected absences—circumstances in which reduced activity was entirely foreseeable and legitimate. The artificial intelligence adoption scores that factored into layoff selections consequently dropped for employees unable to demonstrate active engagement with company systems during protected leave periods.
Meta's broader termination programme, announced in May, affected approximately 8,000 employees representing roughly 10 percent of the company's global workforce. The company framed the reductions as necessary to sharpen focus on artificial intelligence development and research, positioning the layoffs as a strategic reorientation rather than financial distress. Meta's legal representatives argued during Thursday's hearing that human managers retained ultimate decision-making authority and that the AI systems merely supported their judgments. The company further contended that dismissed workers retained access to health insurance through other means, diminishing the irreparable harm argument, and characterised the typical remedies available in subsequent arbitration as adequate compensation for any discrimination that might be proven.
The plaintiffs' legal team countered that the structural dependence on AI systems fundamentally altered the evaluation calculus for employees, particularly those managing medical conditions, caring for family members, or exercising statutory rights to medical leave. During oral arguments, lawyer Barbara Cowan emphasised to Judge Orrick that certain life circumstances—pregnancy, childbirth, active medical treatment, bonding with newborns—cannot be paused or remedied after the fact, making the immediate termination consequential in ways that transcend financial quantification. The argument invoked a emerging body of employment law scholarship questioning whether traditional remedies in arbitration adequately address dignitary harms and categorical exclusions that algorithmic systems may impose.
The ruling simultaneously affirmed Meta's short-term legal position while leaving substantial uncertainty about the ultimate merits of the discrimination claims. Judge Orrick explicitly stated in his written order that he might reconsider his decision if the plaintiffs presented additional evidence demonstrating how and to what extent artificial intelligence actually shaped individual termination decisions. This language suggests that the court recognises the novelty and seriousness of the underlying allegations even while declining emergency intervention. The judge's apparent openness to modification creates incentives for plaintiffs' lawyers to develop more granular evidence about the company's AI systems during the discovery phase of arbitration proceedings.
The case exposes a significant gap in employment law frameworks designed for previous technological eras. Arbitration agreements now standard across major technology companies typically include exceptions permitting workers to seek temporary restraining orders in situations involving alleged theft of trade secrets or recruitment solicitation, yet these same agreements rarely contemplate emergency relief in connection with layoffs of at-will employees. The plaintiffs argue that algorithmic discrimination in workforce termination constitutes a sufficiently novel and serious harm to justify extending temporary relief exceptions beyond their traditional applications. However, Meta and other large employers counter that at-will employment principles have long permitted dismissals without cause, and that the introduction of AI technologies does not fundamentally alter that legal framework.
The terminations have already substantially progressed. Meta removed dismissed workers from company systems on May 20 following the initial notification letters, and they have not performed work since that date, though they technically remained on payroll pending the legal proceedings' resolution. Many workers face final termination on July 22, with others experiencing finalisation of their departures during late July or August depending on their location and contract terms. The geographic and temporal staggering of terminations reflects Meta's apparent effort to manage operational disruption while the legal challenges proceed, but it also complicates plaintiffs' efforts to demonstrate collective harm and identify precise causal mechanisms linking AI system outputs to individual dismissal decisions.
For Malaysian and Southeast Asian readers, this case carries significant implications as technology companies expand their regional operations and employment bases throughout the region. The absence of clear regulatory frameworks governing algorithmic decision-making in human resources creates vulnerability for workers across jurisdictions with limited precedent on artificial intelligence accountability. Should Meta's attempted layoffs proceed successfully despite pending litigation in the United States, the company may feel emboldened to deploy similar AI-based termination methodologies in other countries, potentially including Malaysia, where employment protections may prove even less developed than in California. The outcome of the arbitration proceedings and any subsequent appellate review will substantially influence how technology companies across Asia approach workforce reduction decisions in coming years.
The plaintiffs' motion for a preliminary injunction—a longer-lasting temporary order that could extend relief beyond the immediate emergency phase—remains pending before Judge Orrick. This avenue may ultimately prove more consequential than Friday's emergency ruling, as it would permit the judge to halt layoffs on a more sustained basis while the full merits of the discrimination claims proceed through arbitration. The preliminary injunction standard requires a different analysis than emergency restraining orders, focusing on the likelihood of success on the merits, the balance of hardships, and the public interest in the litigation's outcome. Judge Orrick's hint that he might reconsider his determinations suggests the preliminary injunction motion could receive more sympathetic consideration, particularly if plaintiffs develop evidence that AI systems operated in demonstrably discriminatory ways.
