Meta is defending itself against a lawsuit filed by 26 current employees who allege the technology giant systematically used artificial intelligence to identify and eliminate workers on protected leave during a major restructuring exercise in May. The case, filed in federal court in Oakland, California, centres on Meta's decision to cut 8,000 employees—roughly 10 per cent of its workforce—and raises significant questions about how algorithmic systems can inadvertently or deliberately discriminate against vulnerable worker populations, an issue with particular relevance for Southeast Asia's rapidly expanding tech sector.
According to the complaint, Meta employed a constellation of algorithmic tools to determine whom to lay off, including keystroke monitoring, activity-tracking systems, AI token-usage metrics, and performance rankings generated through algorithmic assistance. The lawsuit argues that these methods systematically disadvantaged employees who were absent from work due to pregnancy, parental responsibilities, medical treatment, or disability accommodations. Because workers on protected leave inherently generate lower measurable output during their absence, the AI systems flagged them as lower performers, the plaintiffs claim, despite legal protections explicitly designed to shield such workers from employment consequences.
The demographic composition of the plaintiff group underscores the gendered dimensions of the allegation. Roughly half the employees suing took caregiving or pregnancy-related leave, including eight women who took maternity leave, four men who took parental leave, and one woman who took bereavement and family care leave. These workers represent a cross-section of Meta's workforce, yet they argue they were systematically selected for termination through a process ostensibly designed to be objective and meritocratic. One male plaintiff alleged that his manager actively discouraged him from taking approved medical leave for a serious health condition, explicitly warning that doing so would render him vulnerable to the anticipated layoffs—a claim that suggests discriminatory intent may have extended beyond algorithmic selection to human management decisions.
Meta's response to the allegations has been characteristically blunt. The company stated that the claims "lack merit and are not based on facts," insisting that "workforce management and organisational decisions were and are made by people, not AI." This assertion sidesteps the core legal argument: that even if people made the final decisions, they relied on algorithmic rankings and performance metrics that were inherently biased against those on protected leave. The lawsuit does not argue that Meta intentionally programmed discrimination into its systems; rather, it contends that Meta failed to adjust its metrics to account for legally protected absences, creating a facially neutral process with discriminatory outcomes.
The legal theories underpinning the case span multiple federal and state statutes. The 26 employees invoke the Family and Medical Leave Act, which guarantees unpaid, job-protected leave for qualifying medical and family reasons; the Americans with Disabilities Act, which requires reasonable accommodations for disabled workers; the Pregnancy Discrimination Act; and the Pregnant Workers Fairness Act. Collectively, these laws establish that employers cannot penalise workers for exercising protected rights, nor can they use neutral policies that have a disproportionate impact on protected classes without demonstrating those policies are necessary for the job.
The "disparate impact" doctrine at the heart of the case has become unexpectedly contentious in the Trump administration's second term. The concept, embedded in Title VII of the Civil Rights Act of 1964, holds that facially neutral employment practices can be unlawful if they disproportionately burden a protected class—in this instance, women and disabled workers—unless the employer can demonstrate the practice is necessary. The Trump administration has explicitly moved to deprioritise disparate impact enforcement, arguing it undermines meritocracy and rests on a false assumption that workforce imbalances inevitably reflect discrimination. This ideological shift has already influenced the Equal Employment Opportunity Commission's enforcement posture, leading it to drop some discrimination cases.
However, the Meta lawsuit reveals a significant limitation to the Trump administration's enforcement retreat: companies remain vulnerable to private litigation even if the government deprioritises enforcement. Workers can pursue disparate impact claims independently if the EEOC declines to act on their behalf, and several states—including California, where Meta's case was filed—have enacted laws that explicitly prohibit disparate impact discrimination. This creates a bifurcated enforcement landscape in which federal enforcement may weaken while state and private litigation remains robust, potentially leaving tech companies more exposed to costly suits than they might expect.
For Malaysian and Southeast Asian stakeholders, the Meta case illuminates a critical emerging challenge in the region's technology sector. As artificial intelligence increasingly powers workforce management systems across multinational technology companies with operations in Malaysia, Singapore, the Philippines, and beyond, the question of algorithmic fairness becomes pressing. Many Southeast Asian countries lack sophisticated employment law frameworks to address algorithmic discrimination, and their regulatory bodies have limited capacity to scrutinise how AI systems are deployed in hiring, promotion, and termination decisions. The Meta case provides a cautionary template for how seemingly objective systems can encode existing biases or create new forms of structural discrimination.
The plaintiffs' lawyers have made a strategic request: they are seeking to preserve the status quo by keeping the 26 workers employed pending arbitration. This framing emphasises the irreversibility of the harm. Once separations become final, the damage extends beyond lost wages. Workers lose employer-subsidised health coverage during pregnancy and postpartum recovery periods, forfeit time-bound family leave rights, lose unvested equity, and potentially face immigration consequences if they held work visas. These cascading harms underscore why employment discrimination law prohibits penalising protected leave in the first place—the consequences ripple far beyond a single paycheck.
The technical specificity of Meta's approach also matters. Rather than relying on subjective manager assessments, Meta attempted to ground its selection process in measurable, algorithmic metrics—keystroke activity, code token usage, and other quantifiable outputs. This technical approach was likely intended to insulate the company from bias accusations, yet it may ultimately deepen the legal vulnerability. By automating the process without human oversight of whether the metrics fairly accounted for legally protected circumstances, Meta may have created a more defensible claim of systemic discrimination than had managers simply chosen arbitrarily. The lawsuit thus presents a paradox: the very technological sophistication Meta deployed to ensure fairness may have instead systematised the unfairness it failed to prevent.
As the case proceeds, it will generate significant precedent for how courts evaluate algorithmic employment decisions under anti-discrimination law. If the plaintiffs prevail, companies will face pressure to audit their AI systems for disparate impact before deployment, to build in explicit accommodations for protected leave, and to retain human review mechanisms that can catch algorithmic bias. Conversely, if Meta succeeds in demonstrating that its systems were properly designed and that personnel decisions ultimately rested with humans, the ruling could establish that companies can use sophisticated monitoring and algorithmic ranking without incurring liability, provided they disclaim algorithmic determinism in their defence. Either outcome will reverberate through global technology companies, including those hiring and managing workers across Southeast Asia.
