Former employees sue over Meta's alleged use of biased AI systems during layoffs

Twenty-six former Meta employees are suing the company for allegedly using biased AI tools that "disproportionately selected" people who took medical leave as candidates to be laid off, Reuters reports. All the plaintiffs in the lawsuit were impacted by Meta's most recent round of layoffs, which cut around 8,000 employees in the hopes of offsetting the company's investments in AI and data center infrastructure.

Meta is accused of using "a constellation of artificial intelligence systems" to select which employees should be cut. It's claimed those included an internal AI assistant called "Metamate," employee-trained "second brain" agents, AI-token usage dashboards and "keystroke- and activity-monitoring data." The alleged goal was to identify and rank employees based on things like their performance and productivity, but also how "AI-native" they are and how many AI tokens they use. The problem, the lawsuit claims, is that Meta's system didn't take into account employees on family or medical leave, or whose disability might affect their output, despite them being in situations where they couldn't use more tokens or engage with other AI tools.

Engadget has asked Meta to comment on the lawsuit and its use of AI. In a statement given to Reuters, the company said the suit lacks merit as the layoffs always involved human input. "Workforce management and organizational decisions were and are made by ​people, not AI," Meta said.

Key to the lawsuit's argument, each of the plaintiffs "took, requested, or was approved to take statutorily protected leave; attempted to take protected leave and suffered interference; or requested or received a reasonable accommodation for a disability," within 24 months of being laid off. Laws like the Family and Medical Leave Act expressly prohibit companies from considering employees taking protected leave as part of their employment decisions. At the state level, the California Family Rights Act also prohibits it, and the state's Fair Employment and Housing Act "forbids the use of an automated-decision system that produces disparate-impact discrimination on the basis of disability or sex, including pregnancy."

The former employees are asking the court to block Meta from completing its layoffs "pending an independent audit of the algorithmically assisted selection process." Because of the terms of Meta's employment contract, the plaintiffs also intend to pursue their claims in arbitration.

Reuters first reported that Meta was recording its staff's keystrokes, mouse movements and clicks to train AI in April 2026. Besides sounding invasive and creepy, the program reportedly violated the European Union's privacy laws. Meta chose to pause the AI training program when it became clear the private chats and transcripts the system recorded were accessible to employees. Considering the indignity of the company's original plan, the idea that similar surveillance data may have been used to conduct layoffs isn't particularly surprising.