Dozens of employees at Meta have filed a lawsuit against the company, accusing it of using artificial intelligence techniques to identify employees who will be laid off.
The lawsuit, filed Monday in a federal court in California, concerns the layoff of about 8,000 employees by Meta (the parent company of Facebook, Instagram and WhatsApp) earlier this year.
Employees allege that the company used a suite of internal artificial intelligence systems, including automated performance evaluations and data monitoring clicks and computer activity, to identify those who would be fired.
The employees indicated that these tools targeted them after they had requested legally protected leave, maternity leave, or disability accommodations.
The complaint states that the company did not rely on the judgment of managers who knew their employees, but rather used artificial intelligence to automatically evaluate, rank, and select workers.
The 26 plaintiffs listed in the lawsuit are asking the court to issue a preliminary injunction preventing Meta from finalizing layoffs while the case is being considered, in addition to compensation that could include rehiring, back pay, stock options, and other benefits.
The issue of artificial intelligence in the workplace has become controversial, with employees fearing bias and privacy violations, while regulators monitor the legality of these tools. Several US states have recently passed laws to protect workers from automated decision-making systems.
The lawsuit explains that Meta's tools collect data on employee performance and productivity, but this data is incomplete when an employee is on medical or family leave. For employees with disabilities, these metrics are already low, meaning that employees who took protected leave were unfairly targeted for dismissal, as if they were being punished for exercising their legal rights.
The impactful cases include: a scientist who was informed of her dismissal just two days before giving birth, an engineer who received a low rating due to leave for injury, and a class manager who was dismissed 16 days after starting his medical leave.
Meta, for its part, strongly denied these claims, and a spokesperson confirmed that workforce decisions are made by humans, not artificial intelligence.
Meta had launched an AI-powered monitoring program earlier this year to record employee clicks, activity, messages, and location data, with the goal of training its AI systems on employee behavior. Mark Zuckerberg said that the average intelligence of the company's employees is above average, and that the models learn by observing intelligent people at work.
But the company quietly launched the program, via an internal memo from an engineer rather than a senior executive, and without explicitly seeking employee approval, sparking widespread outrage. More than 1,600 employees signed a petition protesting the program, which led Zuckerberg to suspend it last June.
In Monday's lawsuit, the plaintiffs' lawyers asked the court to conduct an independent audit of Meta's tools to determine why the 26 employees, who were on leave or had approved benefits, were terminated. They accused the company of deliberately concealing its employee selection process.
The plaintiffs remain officially employed by the company until July 22, when the layoffs are scheduled to take effect. Their lawyers are asking the judge to allow them to remain anonymous and preserve their employment status during the arbitration, warning that the final termination will result in irreparable losses, such as the loss of maternity and medical insurance, the forfeiture of vacation pay, the loss of stock options, and serious repercussions for their immigration status.
