A group of Meta employees has initiated legal action against the company, alleging the use of artificial intelligence tools to identify staff for significant layoffs. The employees assert that these AI systems disproportionately targeted them after they requested protected or maternity leave or accommodations for disabilities.
The lawsuit, presented on Monday in a federal court located in Northern California, refers to Meta's decision earlier this year to lay off around 8,000 workers. Meta, the parent company of Facebook, Instagram, and WhatsApp, is accused of employing a “constellation of internal artificial intelligence systems,” which included AI performance evaluations and monitoring of keystroke and activity data, to determine who would be dismissed.
The comprehensive complaint, spanning 71 pages, states, “Meta did not compile the termination list through the careful consideration of managers familiar with employee roles.” Instead, the 26 plaintiffs allege that the selection of layoffs was predominantly driven by AI systems that scored, ranked, and ultimately decided employee terminations.
The plaintiffs are seeking an urgent court order to halt Meta from proceeding with the layoffs while their case is underway. They are requesting remedies that might encompass reinstatement, back pay, stock options, benefits, and additional damages.
Concerns surrounding AI decision-making in workplaces are escalating as employees question issues like bias, privacy, and trust, with regulators beginning to examine the legality of these technologies. In recent years, states such as California, Colorado, and Illinois have enacted laws aimed at shielding workers from AI-related biases and the implications of “automated decision systems.”
The litigation contends that Meta's AI tools compile various metrics related to employee performance, productivity, and more. These metrics, however, often lack data on individuals who are on medical or family leave, particularly affecting employees with disabilities, thereby skewing evaluations.
“The consequence was that staff members who utilized their rights to protected leaves were unfairly targeted for layoffs, based on evaluations that not only overlooked their leave status but effectively penalized them for asserting their legal entitlements,” the complaint states.
Among the plaintiffs is a scientist on maternity leave, who received her layoff notice just two days before giving birth. Another plaintiff, an engineer, reported receiving a diminished performance rating due to time off for an injury, while a manager on medical leave stated he was terminated just 16 days into his leave.
In response to these allegations, a Meta spokesperson described the claims as baseless, asserting in an email to The Guardian that “decisions related to workforce management and organization were made by people, not AI.”
Meta launched its AI monitoring initiative earlier this year to track workers’ keystrokes, mouse movements, browsing habits, messages, emails, and location data on company devices. CEO Mark Zuckerberg explained that the goal was to enhance the company’s AI systems by analyzing employee behavior.
“The AI models learn by observing exceptionally talented individuals perform tasks,” Zuckerberg remarked during an internal meeting, according to reports from The Information. “The average intelligence of our team is significantly above the typical level found in other task performer pools.”
The lawsuit accuses Meta of having introduced this monitoring program without adequately informing employees or seeking their consent, as it was communicated through a “low-visibility internal post made by an engineer rather than a senior leader.” The complaint further alleges that some teams received no notification or possibility to opt out.
Following employee pushback, which included over 1,600 signatures on a petition asserting privacy violations, Zuckerberg announced a halt to the program in June.
The plaintiffs' legal team is now requesting the court approve an independent audit of Meta’s AI tools to shed light on why the 26 employees, who were on leave or receiving disability accommodations, were chosen for termination. They assert that Meta has deliberately kept the details of its selection process confidential from its workforce.
Currently, the plaintiffs remain employed by Meta until July 22, when their layoffs are set to commence. Due to concerns regarding potential retaliation, the attorneys are asking the court for permission to keep the identities of the plaintiffs confidential and to maintain their employment status while arbitration processes are ongoing.
The attorneys emphasized, “Once these separations are finalized, the damages are irrevocable: loss of employer-sponsored health benefits during pregnancy, postpartum recovery, and ongoing medical treatment; expiration of time-sensitive leave rights; forfeiture of unvested equity; and triggering of immigration complications.”
