A federal judge in California has declined to prevent Meta from proceeding with planned layoffs affecting 26 employees who contend they were singled out for redundancy through the company's artificial intelligence systems. US District Judge William Orrick in Oakland ruled on Friday that the workers failed to establish the level of urgency and irreparable damage required for him to issue an emergency halt, allowing Meta to execute the terminations beginning July 22 as scheduled.
The lawsuit, believed to be unprecedented in targeting a major technology firm's use of AI during workforce reductions, alleges that Meta deployed algorithmic tools to identify candidates for elimination based on metrics that inherently disadvantaged those dealing with medical issues or family care responsibilities. The workers claim that these systems, which measured productivity scores and artificial intelligence adoption rates, systematically penalised absences due to legitimate health conditions or protected leave periods, creating a discriminatory effect against disabled employees and those requiring medical treatment.
Meta has consistently rejected allegations of misconduct and maintains that human personnel made all final decisions regarding which positions would be eliminated. The company did not respond to requests for additional comment following the court's decision. However, the judge's written ruling left the door open for future reconsideration, explicitly noting that he might alter his position if the plaintiffs present additional evidence demonstrating how artificial intelligence factored into the reduction in force process.
The technological systems at the centre of this dispute paint a troubling picture of workplace surveillance and algorithmic decision-making. According to court documents, Meta relied on multiple internal AI systems when compiling termination lists, including a large language model assistant called Metamate designed to function as an employee-trained "second brain" that monitored workers' communications and digital documents. Additionally, a productivity scoring mechanism scanned employee keystrokes, screen activity, email exchanges and internet history to generate performance metrics that fed into layoff determinations.
The workers' legal representatives emphasised a critical flaw in Meta's implementation: these monitoring systems continued operating throughout vacation periods and legally mandated leave times, causing workers' artificial intelligence adoption scores to decline precisely when they were entitled to be absent. This mechanical approach to performance assessment meant that employees following the law in taking protected leave were penalised during the algorithm's evaluation phase, creating what amounts to a built-in bias against legally protected absences.
When Judge Orrick considered whether to block the layoffs, he weighed whether the 26 workers had demonstrated "irreparable harm," the legal threshold necessary for emergency judicial intervention. The defence argued that any financial losses, including forfeited salaries and employer-provided benefits, could theoretically be recovered through monetary damages if the workers ultimately prevail in arbitration proceedings. The plaintiffs' legal team countered that certain damages simply cannot be remedied through compensation after the fact, pointing to the loss of health insurance coverage and the accompanying disruption to critical medical treatments.
Barring workers from health insurance represents an especially acute concern for those managing pregnancies, ongoing medical conditions, or serious treatments where continuous coverage proves essential. During Thursday's hearing, lawyer Barbara Cowan argued forcefully that some aspects of employment termination cannot be undone, no matter what financial settlement might later be awarded. The loss of bonding time with newborns, pregnancy interruption, or delays to active medical care inflict harms that money cannot fully repair, she contended.
Meta's counsel countered that terminated workers were merely losing employer-subsidised insurance options rather than losing coverage entirely, characterising this as a standard type of economic harm that arbitration could adequately address through financial remedies. This framing underscores a fundamental dispute about whether employment law adequately protects workers when algorithmic systems make decisions about who stays and who goes, particularly when those decisions may disproportionately affect protected classes.
The broader context here involves the tension between mandatory arbitration clauses and emergency relief exceptions. Meta's employment agreements require workers to resolve disputes individually through private arbitration rather than pursuing collective action through conventional courts. However, arbitration agreements typically include carve-outs allowing workers to seek temporary restraining orders and preliminary injunctions in court while their underlying claims proceed through arbitration. While such exceptions exist, they have traditionally applied in cases involving alleged theft of business secrets or poaching of employees, not mass layoffs of at-will workers, making this case's legal arguments genuinely novel.
The May layoffs affected approximately 8,000 Meta employees globally, representing roughly 10 percent of the company's total workforce. Meta initiated this reduction as part of a strategic pivot toward expanding artificial intelligence capabilities and investment. The 26 workers filing suit include engineers, managers, researchers and designers who remain formally on the payroll but have been locked out of company systems since May 20 and have performed no work since that date. Final termination for many occurs on July 22, with additional separations scheduled for later in July or August.
For Malaysian and Southeast Asian observers, this case carries significant implications as technology companies throughout the region increasingly adopt artificial intelligence systems for human resources functions. The lawsuit exposes how algorithmic decision-making in layoffs can embed discrimination against protected groups, even when no conscious bias from individual managers exists. As Malaysian firms consider deploying similar technologies, this precedent suggests that algorithmic systems used for employment decisions must be carefully audited to ensure they do not perpetuate or amplify discrimination against employees with disabilities or those requiring medical leave.
The judge's suggestion that he might reconsider his ruling with additional evidence indicates the legal landscape around AI-driven employment decisions remains fluid and contested. The plaintiffs' preliminary injunction motion remains pending, meaning further developments could alter Meta's timeline or enforcement procedures. Ultimately, whether through arbitration or other legal channels, this case will likely establish important precedent governing how companies throughout the technology sector and beyond can legitimately deploy artificial intelligence in workforce reduction decisions without violating anti-discrimination protections.
The case fundamentally challenges assumptions about artificial neutrality in algorithmic systems. While AI tools are often presented as objective alternatives to human bias, this lawsuit demonstrates that algorithms trained on historical data or designed without considering protected categories can generate discriminatory outcomes as effectively as subjective human decisions. For technology companies expanding AI usage across operational functions, including human resources, the Meta case serves as a cautionary reminder that good intentions and claims of algorithmic neutrality do not substitute for rigorous testing to ensure systems comply with employment discrimination law.
