A flashing cursor in an empty text editor is silent. It does not type. It does not request API tokens, trigger keystroke trackers, or update project dashboards. To the automated systems tracking productivity at Meta, a silent computer is a dead computer. And a dead computer belongs to an employee who, on paper, is contributing nothing.
This is the silent engine at the heart of a federal lawsuit filed in Oakland, California. Twenty-six Meta employees are suing the social media giant. They claim that when the company decided to slash about ten percent of its workforce—roughly 8,000 people—it didn't rely on the human judgment of managers who knew their teams. Instead, they argue, Meta handed the scissors to a cold constellation of internal algorithms, keystroke monitors, and performance-ranking software.
The problem with an algorithm is that it lacks context. It only knows what it can count. If you are on parental leave, recovering from childbirth, or managing a severe medical crisis, your keystroke count is zero. Your token usage is zero. The system does not see a human being taking legally protected time off. It sees a flatlining line graph.
Consider the experience of a scientist who was on approved leave, preparing to bring a new life into the world. Her focus was where it should have been: on her impending labor and postpartum recovery. But the tracking software was still running, logging her absence not as a human right, but as a performance failure. She was notified of her termination just two days before she gave birth.
The Math of Being Human
When we look at workforce data, numbers seem objective. But the way those numbers are collected is deeply biased. If an automated system evaluates employees by stacking their recent activity scores against one another, any gap in activity becomes a vulnerability.
The lawsuit alleges that Meta’s automated ranking systems are designed in a way that makes it impossible for anyone on family or medical leave to maintain a competitive score. One software engineer covered by the suit took time off to recover from a physical injury. He returned to find his internal rating dragged down by the weeks he spent healing. Another manager had been on approved medical leave for just sixteen days before the system flagged him for termination.
And then there are the warnings. One employee, struggling with a serious health condition that Meta’s own medical provider had validated, was warned by his manager not to take his approved leave. The manager’s warning was simple: if you step away, the system will select you for the upcoming cuts.
He was trapped. Work through a disability, or take the time to heal and watch an automated system delete your career.
Meta has denied the allegations. A spokesperson stated that the claims lack merit, asserting that workforce decisions "were and are made by people, not AI."
But the workers argue that even if a human hand ultimately clicked the "approve" button on the layoff list, the lists themselves were generated by a process heavily weighted by automated tracking. When humans rely blindly on algorithmic recommendations, the machine is the one making the choice.
The Inequity of the Tracker
The impact of this automated evaluation does not fall equally. Statistically, women are far more likely to take pregnancy, maternity, and caregiving leave. By building a system that automatically penalizes prolonged gaps in digital activity, the algorithms disproportionately target women.
Of the twenty-six plaintiffs, about half took leave for caregiving or pregnancy. Eight are women who took maternity leave. Four are men who took parental leave to care for their newborns. Another is a woman who stepped away to care for a sick relative, only to have that leave transition into bereavement leave when they passed away.
These are the moments that define a human life. They are the moments where we are most vulnerable, and where we most rely on the safety nets built into labor laws. The Family and Medical Leave Act, the Americans with Disabilities Act, and the Pregnant Workers Fairness Act exist specifically to protect people during these times. They are meant to ensure that choosing to care for a dying parent or a newborn child does not cost you your livelihood.
When automated tracking dashboards are allowed to run without a manual override, they effectively bypass these legal protections. The software does not pause to consider whether an absence is legally protected. It simply registers a drop in output.
The Cliff's Edge
The 26 employees are still technically employed by Meta, but their official separation is scheduled to begin on July 22. Because of this, their lawyers are rushing to secure an emergency injunction to halt the layoffs and keep them on the payroll while the dispute goes to arbitration.
They are fighting against a clock, and the stakes are incredibly concrete.
For the scientist who just gave birth, the loss of employment means the immediate loss of her company-subsidized health insurance during postpartum recovery. For others, it means the sudden forfeiture of unvested stock options they worked years to earn, or the immediate threat of deportation as their work visas expire.
These are not minor inconveniences. They are life-altering disruptions.
When we let algorithms run our workplaces, we trade human empathy for administrative speed. A line manager might know that an engineer is away because their child is in the hospital. A manager might remember the late nights that engineer put in six months ago. But to a keystroke logger, only today's silence matters.
The legal battle in Oakland is about much more than Meta. It is a trial run for the future of work. We have to decide if we are willing to let our lives be reduced to data points, evaluated by systems that can measure every single click of a mouse but cannot comprehend the value of a human life.
