Meta used AI workplace tools to target employees on medical leave, lawsuit alleges

• Twenty-six employees sued Meta, saying it used AI systems to help select workers for layoffs.
• The lawsuit says AI usage metrics penalized workers who took medical, parental, or disability leave.
• A Meta spokesperson said that the “claims lack merit and are not based on facts.”

Meta used AI-powered workplace systems to penalize employees for taking medical and parental leave and then selected them for layoffs, according to allegations in a new lawsuit.

The complaint, filed by 26 current and former workers on Monday in federal court in Northern California, alleges that Meta relied on AI-assisted tools and employee-monitoring data to score and rank workers before cutting about 8,000 jobs in May.

The plaintiffs allege that the process relied on metrics such as work output, software development activity, and AI tool usage that employees could not accrue while on protected leave. By failing to adjust those measurements for time away, the suit alleges, Meta effectively recorded medical leave, maternity leave, and disability-related reductions in output as underperformance.

“Meta did not assemble the termination list through the considered judgment of managers who knew the work,” the complaint says.

Instead, the employees allege, the company relied on systems including its internal AI chatbot, Metamate; AI usage dashboards; employee-trained “second-brain” agents designed to reproduce parts of a worker’s output; activity-monitoring data; and AI-assisted performance and calibration tools.

They allege that Meta ranked employees using internal AI-adoption categories such as “AI Native,” “AI First,” and “AI Enabled,” and that employees’ scores declined when they were away from work. The allegations have not been tested in court.

“These claims lack merit and are not based on facts,” a Meta spokesperson told Business Insider. “Workforce management and organizational decisions were and are made by people, not AI.”

The case comes as Meta has pushed employees to use AI more heavily and tracked adoption through internal dashboards and leaderboards. Other major companies, including Disney, JPMorgan, and Visa, have also begun measuring employee AI usage.

The lawsuit against Meta referenced internal AI programs the company introduced this year. Meta had installed software on many US employees’ computers to capture keystrokes and mouse movements as training data for its AI models, prompting employee protests over privacy and the program’s mandatory nature, Business Insider reported in April. The lawsuit alleges that data from this monitoring program also helped feed Meta’s layoff-selection systems.

Meta paused the program in June after an internal leak made the data widely accessible across the company.

Meta also introduced a performance system called Checkpoint this year, placing greater emphasis on outcomes and giving the company’s top-rated workers significantly larger bonuses. The lawsuit alleges that an AI-enabled element of Checkpoint made employee AI adoption a “core assessment metric.”

The complaint says several employees were selected for layoffs while on approved leave or shortly after returning from it. One engineer alleged that his manager blamed a lower rating on the “broken time” caused by an injury that prevented him from working. Another employee says a manager warned that taking medically approved leave would lead senior leadership to “definitely” nominate him for layoffs.

The employees are asking a judge to pause their terminations while their claims proceed in arbitration and to order an independent audit of Meta’s layoff selection process. They also want Meta to recalculate the selections without counting protected leave or disability accommodations against workers.

The plaintiffs, who have chosen to remain anonymous, worked at Meta across several states, including California, Washington, New York, Illinois, Pennsylvania, and Florida.

Reuters first reported on the lawsuit on Tuesday.

Have a tip? Contact Pranav Dixit via email at pranavdixit@protonmail.com or Signal at 1-408-905-9124. Use a personal email address and a nonwork device; here’s our guide to sharing information securely.