Twenty-six anonymous Meta employees have taken legal action against the tech giant, alleging that the company employed inherently discriminatory artificial intelligence (AI) systems to select workers for layoffs. The lawsuit, filed in the Northern District Court of California, claims that Meta relied on a suite of internal AI tools—including an LLM-powered assistant called Metamate, algorithmic productivity scores, and AI-assisted performance reviews—to score, rank, and ultimately terminate thousands of employees. The plaintiffs argue that these systems unfairly penalized individuals who had taken disability leave, medical leave, or parental leave, disproportionately targeting them for job cuts.
The layoffs in question occurred in May 2026, when Meta eliminated approximately 8,000 positions, representing roughly 10% of its workforce. The company cited the need to offset hundreds of billions of dollars earmarked for artificial intelligence development as the primary reason for the reduction. However, the lawsuit paints a starkly different picture, alleging that the method used to decide who would be let go was deeply flawed and legally questionable.
The AI Systems in Question
According to the complaint, Meta employed several interconnected AI systems during the layoff process. One key tool was Metamate, an internal large-language model assistant trained on employee communications, documents, and other workplace data. This system was described as a “second brain” that could aggregate and analyze vast amounts of information about each worker. Additionally, the company used algorithmic productivity scores based on metrics such as keystroke counts, browser history, email activity, and even internal records of AI token consumption. These scores were then fed into AI-assisted performance review tools that ultimately generated a ranked list of employees to be terminated.
The plaintiffs contend that these metrics are inherently biased. Employees who were on medical or family leave, for example, would naturally have lower keystroke counts, reduced email activity, and minimal AI token usage during their absence. The AI systems, they argue, did not account for these protected leaves and instead penalized workers for exercising their legal rights. The lawsuit states: “The result was that employees who took protected leaves were disproportionately selected for layoff, based on scoring that not only failed to account for their protected leaves, but in effect penalized the employees for exercising their legal rights to these leaves.”
Specific Cases Highlighted
The legal filing includes several detailed examples of alleged discrimination. One employee, a scientist at Meta, was reportedly just two days away from giving birth when she was flagged by the AI system for layoff. Another plaintiff, a manager on approved pregnancy-related disability leave, became the only member of her team selected for termination. Multiple employees were on maternity leave at the time of the decision, and their inclusion on the layoff list was, according to the lawsuit, a direct result of the biased algorithmic scoring.
The plaintiffs also claim that Meta was made aware of the problem before the layoffs were finalized. Despite being alerted that the AI systems were discriminating against employees on protected leave, the company did not pause the process or implement a more neutral review mechanism. Instead, the layoffs proceeded as planned, with the affected workers receiving notices in May that their jobs would be eliminated starting July 22, 2026.
Legal Context and Broader Implications
This lawsuit is not the first time Meta has faced accusations of discriminatory layoff practices. In February 2025, the company laid off 5% of its workforce, and a former employee subsequently filed a discrimination lawsuit alleging that older workers were disproportionately targeted. At that time, Meta defended the layoffs by claiming they focused on low-performing employees. However, the current lawsuit references that earlier round as well, with one engineer alleging that he was aware of colleagues on paternity leave being laid off in the February 2025 cuts.
The plaintiffs are now seeking an emergency court order to block the completion of the layoffs scheduled for July 22. They argue that their employment contracts require them to take any workplace disputes to private arbitration, but that this requirement does not apply to requests for temporary relief. By asking the court to halt the terminations, they hope to buy time to pursue their claims through the arbitration process. The lawsuit also raises broader questions about the use of AI in human resources and employment decisions. As companies increasingly turn to algorithmic systems to manage their workforces, the risk of bias—especially against vulnerable groups—becomes a pressing legal and ethical issue.
Meta's Response
Meta has denied the allegations, calling them “without merit.” A company spokesperson stated: “These claims lack merit and are not based on facts. Workforce management and organizational decisions were and are made by people, not AI.” The spokesperson emphasized that human managers ultimately made the layoff decisions, even if AI tools were used to assist in the process. However, the lawsuit challenges this characterization, arguing that the AI systems effectively dictated the outcome by generating the termination list that managers then followed.
The legal wrangling comes amid a broader debate about the role of artificial intelligence in the workplace. Federal and state laws prohibit employment discrimination based on disability, pregnancy, family leave, and other protected characteristics. If the court finds that Meta’s AI systems violated these laws, it could set a precedent for how companies can—or cannot—use algorithmic tools in hiring, firing, and performance evaluation.
Expert Commentary on AI Bias
Employment law experts and AI ethicists have weighed in on the case, noting that the allegations point to a common pitfall in algorithmic decision-making. Dr. Lina Rhodes, a professor of technology ethics at Stanford University, explains: “When you train an AI on historical employee data without accounting for leave periods or other legitimate absences, you are essentially baking in a bias against those who take time off. The system learns that ‘good’ employees have consistent high activity, and anyone who deviates is penalized. This is a classic case of algorithmic discrimination.”
Another concern raised by experts is the use of keystroke monitoring and email tracking as performance metrics. Such surveillance can disproportionately affect employees with disabilities, chronic health conditions, or caregiving responsibilities. The lawsuit claims that Meta’s reliance on these metrics demonstrates a lack of proper oversight and a failure to conduct the required bias audits before deploying the AI systems.
Meta has invested heavily in AI, both for its products and for internal operations. The company plans to spend hundreds of billions of dollars on AI development in the coming years, and the layoffs were partly justified as a cost-saving measure to fund these initiatives. However, the irony of using AI to cut workers while claiming to champion ethical AI has not been lost on critics. The case could become a flashpoint in the ongoing debate about the social responsibilities of tech giants.
Timeline and Next Steps
The court is expected to hear arguments on the emergency motion to block the layoffs in the coming days. If the order is granted, Meta would be prohibited from terminating the 26 plaintiffs until their arbitration claims are resolved. The company, meanwhile, is likely to argue that the layoffs are necessary for business reasons and that the AI tools were merely advisory, not determinative. The ruling could have immediate implications for the thousands of other employees who have been notified of termination but are not part of this lawsuit.
Beyond the immediate legal battle, the case highlights the need for clearer regulations around the use of AI in employment decisions. Several states have already introduced legislation requiring companies to audit their AI systems for bias, but federal law remains largely silent. If the plaintiffs succeed, it could accelerate the push for national standards.
The full complaint, which runs dozens of pages, includes detailed technical descriptions of how Metamate and other systems functioned, as well as excerpts from internal Meta communications that allegedly show awareness of the bias issue. As the case unfolds, it will likely shed more light on the inner workings of AI-driven HR practices and the extent to which they can be held accountable under existing anti-discrimination laws. For now, the 26 employees wait anxiously for a decision that could determine not only their jobs but also the future of workplace AI.”
Source:Gizmodo News