Twenty-six Meta employees just filed what could become one of the most consequential employment lawsuits of the decade. Their claim: the company used a web of internal AI systems to decide who gets fired, and those systems had a nasty habit of targeting people who took medical, parental, or disability leave.

The federal lawsuit, filed on July 14 in US District Court for the Northern District of California in Oakland, alleges that Meta deployed what plaintiffs describe as a “constellation of internal artificial intelligence systems” to systematically identify workers for mass layoffs. Every single plaintiff had taken or requested protected leave within the prior 24 months.

What the AI allegedly did

The plaintiffs claim Meta’s AI tools relied on productivity scores, keystroke monitoring, and algorithmically assisted performance rankings to flag employees for termination. The machines were watching how fast you type, how much output you produce, and then ranking you against your peers.

The obvious problem? If you took three months of parental leave or were recovering from surgery, your productivity metrics are going to look pretty thin compared to someone who worked every single day. The AI, according to the lawsuit, didn’t adequately account for that.

The plaintiffs argue there was insufficient human oversight baked into the process. Algorithmic rankings, they claim, were treated as something close to gospel rather than one data point among many.

This layoff round is part of Meta’s broader restructuring, with roughly 8,000 positions on the chopping block. That’s approximately 10% of the company’s total workforce.

A judge has already denied the plaintiffs’ request to temporarily halt the layoffs while the case proceeds.

Why this matters beyond Meta

The core legal tension is straightforward but thorny. US employment law prohibits discrimination against workers who exercise their right to medical or family leave under statutes like the Family and Medical Leave Act. But those laws were written for a world where a human manager decided who to fire. When an algorithm makes the call, proving discriminatory intent becomes exponentially harder.

The plaintiffs’ challenge is demonstrating that the AI’s outputs weren’t just correlated with protected leave but were causally linked to it. Meta will almost certainly argue that the systems evaluated objective performance data and that any disparate impact was incidental, not intentional.

The AI doesn’t need to “know” someone took parental leave to penalize them for it. It just needs to see a three-month gap in keystroke data and draw its own conclusions.

What crypto and tech investors should watch

The European Union already has the AI Act imposing transparency requirements on high-risk AI systems, and employment decisions fall squarely in that category.

The denied injunction is telling. Courts are reluctant to intervene in corporate restructuring, even when AI bias allegations are on the table. Companies considering similar AI-driven layoff processes will read that signal clearly: the legal system, for now, is more likely to let the cuts happen first and sort out liability later.

Whether that calculus changes depends entirely on what discovery in this case reveals about Meta’s internal AI systems, and whether the data shows the pattern the plaintiffs allege.