Twenty‑six current and former Meta employees filed a federal lawsuit this week, accusing the company of using a constellation of AI‑assisted tools to disproportionately identify workers with disabilities or those on approved medical and parental leave for termination during its recent round of mass layoffs. The complaint, filed in Oakland, California, seeks an injunction to pause affected terminations and demands an independent audit and recalculation of layoff selections without counting protected leave against employees.
What plaintiffs allege: AI dashboards, token metrics and monitoring data
According to the complaint, Meta relied on automated or AI‑assisted systems — including an internal chatbot known as “Metamate,” AI usage dashboards and employee “second‑brain” agents — together with activity‑monitoring data to score, rank and ultimately select employees for layoffs. The plaintiffs say those systems measured productivity through metrics such as software development activity, keystrokes, mouse movements and “AI token” usage; employees who took medical or family leave naturally accrued fewer tokens and less activity, thereby lowering their algorithmic scores.
Legal claims and relief sought
The suit alleges violations of federal and state laws that prohibit discrimination and retaliation against workers with disabilities and those taking protected leave, arguing that the automated measures functionally penalized legally protected absences. The plaintiffs — who proceed anonymously in the filing — ask the court to block planned terminations while their claims advance and to order an independent audit of Meta’s layoff‑selection process
Meta’s response: people, not algorithms, make the calls
Meta has publicly rejected the allegations, saying the claims lack merit and emphasizing that workforce management and organizational decisions were and are made by people, not AI. A company spokesperson told reporters that human managers, not automated systems, made layoff decisions, and that the complaint mischaracterizes how internal tools were used.
Context: layoffs and the rise of AI in workforce management
The lawsuit follows Meta’s announced plan earlier this year to cut roughly 10% of its global workforce — about 8,000 positions — as the company rebalances spending amid heavy investments in AI and data center infrastructure. In recent years, employers have increasingly adopted AI‑enabled systems for performance analytics, hiring and workforce planning, creating legal and ethical pressure points about accuracy, fairness and statutory protections for employees.
Mechanics at issue: what the plaintiffs call an AI “constellation”
Plaintiffs describe not a single automated firing machine but a constellation of tools and data sources that, in combination, produced performance scores used in calibration and ranking: Metamate; employee‑trained “second brain” agents that replicate parts of workers’ output; dashboards tracking AI token usage; and activity‑monitoring systems that logged keystrokes and other on‑device signals. The complaint asserts that using raw counts or short‑term activity as core indicators failed to account for legally protected leaves of absence and disability accommodations.
Why plaintiffs say the process disadvantaged leave‑takers
Employees who were on maternity, medical or disability leave could not generate the same volume of outputs, token usage, or keystroke activity as colleagues working full‑time, the suit contends; when those raw metrics were fed into scoring and calibration systems, the absence of contextual adjustment caused protected leave to register as underperformance. The plaintiffs argue that such an approach amounts to de facto discrimination because the metrics correlate with protected conditions rather than actual job quality.
Potential legal and regulatory stakes
If courts find that AI‑assisted or algorithmically derived measures were used in a way that discriminated against protected workers, the case could expand legal scrutiny of the workplace use of generative AI, monitoring software and automated performance systems. Plaintiffs request an independent audit and recalculation of the layoff roster excluding metrics tied to protected leave, a remedy that would test whether companies can lawfully rely on large‑scale automated signals when making high‑stakes employment decisions.
Employer defenses and practical difficulties
Meta and other large employers argue that AI tools are decision‑support systems and that human managers retain judgment authority, complicating plaintiffs’ efforts to prove that algorithms were the proximate cause of discriminatory outcomes. Defendants also note the difficulty of proving disparate impact from complex, multi‑factor systems where human calibration and business context interplay with automated signals. Legal analysts warn, however, that transparency failures and lack of accommodations for protected absences could still produce liability under anti‑discrimination statutes.
Broader implications for workplace AI governance
The lawsuit amplifies wider policy questions about fairness, transparency and due process when employers deploy monitoring and AI systems. Regulators and courts are increasingly focused on whether automated systems incorporate bias controls, adjust for protected leaves, and provide meaningful human oversight — and whether employees have access to audits or explanations for adverse employment outcomes.
What to watch next
Key developments to monitor include whether the court grants the requested injunction to pause terminations, whether the plaintiffs’ arbitration claims proceed given contractual employment terms, and whether regulators or advocacy groups pursue parallel inquiries. An independent audit — if ordered — could set precedents about what constitutes adequate review, remediation and transparency when AI‑assisted metrics influence workforce decisions.
Practical takeaways for employers and counsel
Companies using AI in HR should reassess whether metrics capture lawful, job‑related criteria and whether systems are adjusted for protected absences and accommodations; documentation of human decision‑making and clear governance policies for automated tools can be critical in defending against discrimination claims. Independent audits, recordkeeping and advance impact assessments may increasingly be necessary to demonstrate compliance with anti‑discrimination and employment laws.
