A landmark lawsuit filed in a California federal court may force one of tech's biggest employers to answer a question the industry has largely avoided: what happens when an AI system decides who gets fired?
Twenty-six current and former Meta employees — filing anonymously — have sued the company over its May 2026 round of layoffs, in which roughly 8,000 workers (10% of its global workforce) were let go. The plaintiffs allege that Meta deployed a "constellation of internal artificial intelligence systems" to select employees for termination, and that those systems systematically disadvantaged workers who were on approved medical or family leave, or whose output was reduced by a disability.
Meta, for its part, has rejected the claims entirely. A company spokesperson told Reuters and CNBC that the allegations "lack merit and are not based on facts," adding that "workforce management and organizational decisions were and are made by people, not AI."
What the Algorithm Allegedly Measured
The lawsuit's most striking detail is its specificity about the metrics Meta's internal AI tools allegedly relied upon when building termination lists. According to the plaintiffs' legal filing, the system drew on inputs including performance ratings, calibration scores, productivity and output metrics, "AI-native" ratings, and — notably — AI token consumption.
The plaintiffs argue these inputs are structurally incapable of reflecting the work of employees who were not actively on the job. An employee on approved FMLA leave, recovering from surgery, or managing a chronic condition that reduces their throughput would, by design, score poorly on every one of those dimensions — not because they are underperformers, but because they were legally absent.
"Those tools draw on inputs — performance ratings, calibration scores, productivity and output metrics, 'AI-native' ratings, and AI-token consumption — that, by design, cannot be accumulated by an employee who is on protected medical or family leave, or whose output is reduced by a disability," the plaintiffs' lawyers wrote in the legal filing.
The plaintiffs are not only seeking compensation — they are asking a federal court to issue a preliminary ruling barring Meta from fully implementing the layoffs while they pursue the matter in private arbitration and while an independent audit of the algorithmic selection process is conducted. Those affected were informed in May that their roles would be eliminated starting July 22, 2026.
A Legal Framework Under Pressure
The case draws on federal and state laws that prohibit discrimination or retaliation against employees with disabilities, those taking medical leave, and pregnant workers. The fact that the lawsuit was filed in California, with plaintiffs spread across six states including New York, also raises questions about whether Meta adequately tested its AI systems for bias — a requirement that is increasingly codified into law.
California Governor Gavin Newsom signed more than 20 new AI-related laws in recent years, expanding the state's framework to cover automated decision systems, data privacy, and generative AI. New York City has enacted similar measures. Together, these laws create a growing compliance environment specifically targeting algorithmic decision-making in employment — and this lawsuit may become the first major test of how courts interpret those laws when applied to mass layoffs at scale.
This legal battle is part of a broader reckoning for the AI industry explored in our coverage of the publisher lawsuit against Google's Gemini and AI's collision with existing legal frameworks.
Not Meta's First Controversy Around Internal AI Use
The May layoffs weren't the first time Meta's internal use of AI on its own employees drew scrutiny. In April 2026, the company faced significant backlash after it emerged that Meta had installed new tracking software on US-based employees' computers — capturing mouse movements and keystrokes — with the stated intent of using that data to train its AI models. The initiative, known internally as the Model Capability Initiative (MCI), was paused after intense employee pushback over data privacy concerns.
The parallel is uncomfortable: Meta was reportedly capturing the raw behavioral signals of its own workforce as AI training material, and those same employees are now alleging that AI-driven analysis of their productivity signals determined who would be shown the door. Whether or not those two programs are directly connected, the sequence has sharpened internal and external scrutiny of how Meta governs AI within its own walls.
The Broader Implications for AI-Assisted HR
What makes this lawsuit significant beyond Meta itself is the precedent it could establish. Algorithmic performance management — using productivity software, engagement scores, and output metrics to evaluate employees — is now commonplace across the enterprise technology sector. Rarely, however, has a company faced a lawsuit that specifically names the AI input variables used in a mass termination decision and alleges that those variables were structurally biased against protected classes.
If the plaintiffs succeed in forcing an independent audit of Meta's algorithmic selection process, it could establish a template for challenging AI-assisted workforce decisions at any large employer. That prospect has implications well beyond this case — as we examined in our analysis of how enterprises are grappling with AI governance and unintended consequences.
For now, the case raises a question that no amount of corporate denial fully answers: if workforce decisions are truly made by people and not AI, why does the lawsuit's description of the input variables — AI token consumption, AI-native ratings, output metrics — read like a machine learning feature list?
What's Next
The plaintiffs are seeking a preliminary court order to pause the July 22 implementation of the layoffs pending arbitration and an independent audit. Meta has indicated it will contest the claims. Given the California venue, the scale of the affected workforce, and the specificity of the AI bias allegations, this case is likely to move quickly — and the outcome could reshape how companies document, disclose, and defend algorithmic workforce decisions for years to come.
The AI industry's legal exposure continues to widen, as we've tracked in our coverage of AI accountability in practice and the expanding wave of litigation targeting AI systems.
Related Reading
- Hachette, Elsevier, and Cengage Are Suing Google Over Gemini — and the Entire AI Industry Is Watching
- The "Tokenmaxxing" Crisis: How Enterprises Are Burning AI Budgets
- A Fake Photo of a Hospitalized Senator Just Proved Deepfake Detection Actually Works
- Satya Nadella's AI Warning: You're Paying Twice for Enterprise AI