Meta is facing a lawsuit from 26 current and former employees, who claim that the company used artificial intelligence systems to help decide who would be fired, with a disproportionate focus on employees with disabilities and those who took protected leaves. Meta rejects the claims and emphasizes that organizational decisions were and continue to be made by humans, not by artificial intelligence.

The plaintiffs claim that AI-powered tools, identified in the lawsuit as Metamight and Checkpoint, scanned employee data to score, rank and select people for layoffs. According to the lawsuit, performance metrics did not adjust or offset legally protected absences such as medical leave, parental leave, family care or maternity leave, so the absences of employees on protected leave were recorded as reduced performance and not as suspended or neutral activity - an approach that the lawsuit claims naturally harmed women in particular.

Almost half of the plaintiffs claim they were fired close to maternity or parental leave, without the system taking that into account. One case described in the lawsuit involves a Meta scientist who was allegedly selected for firing a day before her water broke during pregnancy (and two days before the birth itself); another case mentions an engineer who, according to him, received a lower rating due to interrupted work following an injury.

The lawsuit describes Meta's internal performance infrastructure as one that combines keyboard keystroke data and activity tracking, dashboards for the use of artificial intelligence tokens, and algorithm-assisted ratings. According to the claims, the Checkpoint system turned the adoption of artificial intelligence by employees into a core evaluation metric.

The plaintiffs argue that the algorithms used to identify candidates for layoffs relied on biased data, particularly in the way productivity, adjustment needs and role criticality were assessed for employees with disabilities. They also claim that the tools did not include a personal and neutral examination of leaves and accommodations as required by law. The layoffs described in the lawsuit are part of a broader wave of reductions: Meta laid off about 8,000 employees, about 10% of its global workforce, in May - in one of the largest rounds of cuts in the company's history.

The plaintiffs accuse Meta of violating federal and state laws prohibiting discrimination or retaliation against employees with disabilities, those taking medical leave or pregnant workers. According to the lawsuit, the company did not update its artificial intelligence systems to be free of bias, contrary to new laws in California and New York City, which require in-depth audits for automated decision-making tools. The case highlights the growing legal criticism of the use of artificial intelligence in recruitment, evaluation and reorganization of personnel, and it also relates to European regulation: the European Union's Artificial Intelligence Act classifies AI systems related to employment as high risk, and imposes strict requirements for transparency and inspection on them.

According to the AP news agency, the plaintiffs asked the court for a preliminary decision to stop the layoffs and preserve the status quo as long as their claims proceed in private arbitration, arguing that the damages from final layoffs would be irreversible. The lawsuit mentions in this regard a previous case of a former employee, Nicola Fransha: According to Fransha's lawsuit, filed last March, Meta's workforce reductions disproportionately affected older employees - employees over the age of 50 were allegedly 2.5 times more likely to be fired compared to younger colleagues.