A federal judge in California has denied an emergency request from 26 Meta Platforms workers seeking to halt planned job terminations while they challenge allegations that artificial intelligence systems were weaponized to identify and remove disabled employees and those who took medical leave. The decision, delivered Friday by U.S. District Judge William Orrick in Oakland, clears the way for Meta to execute layoffs scheduled to begin July 22, even as the workers pursue what may be the first major lawsuit of its kind against a Silicon Valley giant over alleged discriminatory use of AI in workforce reduction.
The crux of the dispute centres on Meta's deployment of internal AI-driven tools during the company's sweeping May restructuring that eliminated roughly 8,000 positions—approximately one-tenth of its global headcount. According to the lawsuit, the technology giant utilised multiple algorithmic systems to score, rank, and ultimately select employees for termination. These included a large language model assistant branded as Metamate, a productivity measurement system that analysed keystrokes, screen activity, emails and browsing patterns, and an employee adoption score measuring how extensively workers engaged with AI tools themselves.
The plaintiffs, who remain anonymous in court filings but represent engineers, managers, researchers and designers across the company, contend that Meta's systems created a structural bias against workers who required time away from their desks for medical reasons or family care responsibilities. Because these employees accumulated lower AI adoption scores and productivity measurements during absences, the algorithmic systems systematically flagged them for elimination. Meta's decision not to pause or adjust these surveillance mechanisms during vacation periods and legally protected leave further disadvantaged vulnerable workers, the suit argues.
Judge Orrick determined that the workers had failed to demonstrate the severe, irreversible harm required to justify an emergency court order halting the layoffs during the arbitration process. He distinguished between the immediate financial consequences of job loss—which he suggested were theoretically recoverable through later compensation if the plaintiffs prevail in their claims—and the type of irreparable injuries courts typically shield through temporary relief. Yet his written order contained a notable opening for the workers, explicitly stating he might reconsider this position once presented with additional evidence about how Meta specifically deployed AI in the reduction in force.
The workers' legal team responded by emphasising this qualification, arguing that the judge had fundamentally acknowledged the legitimacy of their concerns. In a joint statement, the lawyers stressed that Judge Orrick had recognised the lawsuit addresses "serious questions" about Meta's conduct and that the court remained open to revisiting the injunction decision with more detailed information about the company's algorithmic processes. This positioning reframes what appeared to be a judicial loss into a potential strategic advantage for the plaintiffs' ongoing arbitration claims.
The dispute also highlights a tension in American employment law between workplace arbitration clauses and emergency relief provisions. Meta's employment agreements require individual workers to arbitrate disputes rather than pursue collective court actions, a standard practice at major technology and corporate employers. However, the workers argue that arbitration agreements typically contain exceptions allowing parties to seek temporary relief through courts while the underlying dispute proceeds to private arbitration. Meta's attorneys contended such exceptions rarely apply to standard at-will employment terminations, distinguishing them from cases involving theft of intellectual property or employee poaching.
The stakes for affected workers extend far beyond immediate salary loss. During Thursday's hearing, workers' counsel Barbara Cowan emphasised that the terminated employees faced elimination of stock options, loss of employer-subsidized health insurance, and potential gaps in coverage for critical medical needs including pregnancy care and ongoing treatment. Meta's representative countered that workers retained access to insurance through alternative means, minimising the claimed harm. These competing characterisations of financial and medical consequences proved decisive in the judge's rejection of the emergency restraining order.
Meta has declined to comment directly on the lawsuit or the court decision, instead issuing a blanket denial of wrongdoing through company spokespersons and assertions that human judgment remained central to all termination decisions. The company's legal filings emphasised that workers placed on the termination list in May had been removed from company systems on May 20 and performed no work from that date forward, attempting to establish that the remaining individuals were surplus to operational requirements regardless of algorithmic inputs.
The case carries potentially far-reaching implications for how technology companies deploy artificial intelligence in human resources decisions. As firms across the sector accelerate investments in AI systems to automate recruitment, performance evaluation, and workforce management, this litigation may establish precedent for how employment discrimination law applies to algorithmic decision-making. The anonymity of the plaintiffs themselves reflects the power asymmetry at play—individual workers challenging one of the world's largest technology corporations over opaque algorithmic systems that may disadvantage workers with protected characteristics including disabilities and health conditions.
For Malaysian and Southeast Asian technology workers, the case underscores broader concerns about AI governance and worker protections in rapidly digitising economies. Many regional tech hubs and multinational offices adopt employment practices and systems aligned with global corporate headquarters, meaning algorithmic systems developed and deployed in Silicon Valley eventually influence recruitment and performance management decisions affecting workers throughout Asia-Pacific. The outcome of Meta's litigation could inform how Asian labour regulators and courts approach AI-driven employment decisions, particularly as the region accelerates digital transformation across all economic sectors.
The judge's openness to reconsidering his position based on additional evidence suggests this case remains far from settled even at the preliminary injunction stage. The workers' pending motion for a longer-lasting temporary order could trigger more detailed discovery about Meta's AI systems, potentially forcing the company to disclose proprietary algorithmic processes and training data. Such revelations could strengthen the plaintiffs' ability to demonstrate that Meta's systems functioned in discriminatory ways, whether intentionally or through the propagation of bias in training datasets and model design.
Meta's May layoff round represented one of the technology industry's most aggressive workforce reductions in recent years, announced as the company committed to what executives termed a "Year of Efficiency" driven by artificial intelligence expansion. The company's decision to deploy algorithmic tools in selecting which workers to eliminate—rather than relying on conventional performance metrics and manager assessments—marked a notable shift in how Silicon Valley conducted large-scale restructuring. Whether this approach proves legally defensible may ultimately depend on whether Meta can demonstrate that its AI systems operated neutrally with respect to protected employment characteristics, or whether the plaintiffs can show the algorithms replicated or amplified existing organisational biases.
