Meta reportedly used AI to lay off employees: here’s why this would be unthinkable in France

Meta’s big spring clean culminates in court. Monday evening, 26 current and former employees filed a complaint with the federal court in Oakland, California. According to them, the list of the 8,000 positions eliminated in May (10% of the workforce) was not drawn up by managers, but by a battery of artificial intelligence systems.Meta denies this and insists that these decisions were made “by humans, not by AI”.

How would the algorithm have chosen those selected for dismissal?

The 71-page complaint describes a “constellation” of in-house tools. These include Metamate, the internal assistant powered by large language models, and “second brain” agents trained on each employee’s documents and correspondence. Also included are records from the keystroke-monitoring software deployed in the spring, dashboards measuring AI token consumption, and algorithm-assisted performance rankings.

The heart of the allegation lies in how these scores work: built on measured productivity and recorded activity, they function as continuous oversight in which absence counts as zero, including sick leave. Employees on sick leave, parental leave, or with disabilities would therefore, by design, never have had a chance to catch up with their colleagues, and would have been overrepresented on the list. The 26 plaintiffs (anonymous, from six states and Washington) are asking for the layoffs to be frozen, with effect from July 22, while they take their cases to arbitration.

And in Europe, would this be lawful?

The short answer comes down to one legal provision: since 2018, the GDPR has prohibited decisions producing legal effects (a dismissal obviously falls into that category) from being based solely on automated processing.Real human intervention is needed, not merely a click to validate the machine’s output, failing which the bill could reach €20 million or 4% of global revenue.

The AI Act was supposed to add a second safeguard: employee management systems, from recruitment to dismissal, are classified as high-risk under it, with bias testing, human oversight and employee notification. These obligations were supposed to apply on August 2, except that Brussels has just pushed the entire section back to December 2027 (lawmakers’ sense of timing remains an inexact science). The first major trial over algorithmic dismissal is therefore breaking out in the very week that Europe delays its safety net.

👉🏻 Follow the tech news in real time: add 01net to your sources on Google, and subscribe to our WhatsApp channel.

Source: Reuters