Meta employees file lawsuit over use of AI in layoffs

SAN FRANCISCO (CN) — A group of current and former Meta employees sued the company on Monday after the social media giant used artificial intelligence models to select certain employees and reduce its workforce by 10%.

Twenty-six anonymous plaintiffs filed the lawsuit in federal court for the Northern District of California, alleging that Meta began notifying employees of “retrenchments” starting May 20 and used its AI system to target workers who took or applied for protected leave.

“Meta did not base its layoff list on the prudent judgment of management who knew the business,” the plaintiffs wrote in their 71-page complaint. “Instead, Meta scored, ranked, and selected employees for inclusion on the list using a suite of in-house artificial intelligence systems, including a system internally called ‘Metamate,’ a ‘second brain’ agent trained by employees, keystroke and activity monitoring data, an AI token usage dashboard, and algorithm-assisted performance ranking and adjustment.”

Because the AI ​​used performance reviews, productivity, and other outcome metrics, the plaintiffs argue that the system’s design excludes employees who are disengaged from their jobs and don’t have as many metrics to measure compared to other employees.

“As a result, not only were employees who took protected leave unfairly singled out for termination based on scoring that did not take protected leave into account, but employees who exercised their legal rights to these leaves were effectively penalized,” the plaintiffs wrote.

The plaintiffs hold various roles in meth across the United States, including California, Florida, Illinois, New York, Pennsylvania, and Washington state. They say they have something in common. All had taken, applied for, or been approved for protected leave within the past 24 months.

Plaintiffs are seeking to pursue their claims individually in arbitration because Meta requires its employees to sign a mutual arbitration agreement with a class action waiver. They are asking the court for injunctive relief to halt the layoffs and restore their employment status as of May 20th. Their claims include violations of the National Protected Leave Act, the Family and Medical Leave Act, the Pregnancy Discrimination Act, and the Americans with Disabilities Act.

In April, Meta announced plans to lay off about 8,000 employees, or about 10% of its workforce, and “thousands of additional employees will be redeployed to new artificial intelligence efforts as Meta seeks to transform itself into an ‘AI-first’ company,” the plaintiffs said.

Earlier this year, Meta developed a surveillance program that captures keystrokes, screen content, mouse activity, browser history, messages, email and audio, video, and location data on company-issued devices.

“Data collected through employee tracking programs was used to build AI tools, including by new engineering organizations that redeploy employees on a non-voluntary basis,” the plaintiffs wrote.

The plaintiffs say that Meta’s implementation of the monitoring program was “announced through a discreet internal post, written by an engineer rather than a senior leader in a secondary group, rather than through Meta’s official employee notification channels. There was little notice and no consent or click confirmation. In at least some teams, employees received no consent or confirmation prompts at all, and at least initially had no way to opt out.”

The case will be handled by U.S. District Judge William Orrick, who was appointed by Barack Obama.

“These claims are unsubstantiated and not based in fact. Employee management and organizational decisions have been and continue to be made by humans, not AI,” a Meta spokesperson said in a statement.

Lawyers for the plaintiffs did not respond to requests for comment.