Meta faces lawsuit alleging biased AI drove discrimination in May layoffs

Twenty-six former Meta employees have initiated legal proceedings in the US District Court for the Northern District of California, alleging the technology giant used biased artificial intelligence systems to identify staff for termination during a workforce reduction announced in May 2026. The complaint contends that AI scoring mechanisms, including an internal tool referred to as 'Metamate', disproportionately selected individuals on medical or family leave for redundancy.

The lawsuit describes a "constellation of artificial intelligence systems" used to rank employees, comprising the Metamate assistant, employee-trained "second brain" agents, AI-token usage dashboards, and keystroke- and activity-monitoring data. Plaintiffs argue these tools prioritised productivity and "AI-native" status, measured by token usage, without accounting for protected leave or disabilities that may have temporarily reduced an employee's output.

Specific instances cited in the filing include a scientist selected for termination while on approved pre-birth pregnancy leave, alongside others on maternity, paternity, or medical leave for disabilities. The layoffs affected approximately 10 per cent of Meta’s workforce, impacting around 8,000 employees, as announced by Chief People Officer Janelle Gale in May 2026. The reductions were part of a broader strategy to offset investments in AI and data centre infrastructure.

Meta has denied the allegations, asserting that the claims lack merit. In a statement provided to Reuters, the company emphasised that workforce management and organisational decisions were made by humans, not AI. The social media giant maintains that the lawsuit is not based on facts and that human oversight was integral to the selection process.

The plaintiffs are seeking a preliminary injunction to block further layoffs, which are scheduled to commence on 22 July 2026, pending an independent audit of the algorithmically assisted selection process. The legal action relies on violations of the Family and Medical Leave Act, the California Family Rights Act, and the state’s Fair Employment and Housing Act, which prohibits automated decision systems that produce disparate-impact discrimination based on disability or sex.

Due to the terms of their employment contracts, the plaintiffs intend to pursue their claims through arbitration. The case follows earlier revelations in April 2026 that Meta had paused an AI training program recording staff keystrokes and mouse movements after it was discovered that private chats were accessible to employees, a practice that violated EU privacy laws.