Meta Faces Lawsuit Over Claims It Used AI to Target Disabled Workers in Layoffs

Meta is facing a US federal lawsuit from 26 employees who allege the company used AI-assisted systems that disproportionately selected workers with disabilities, medical conditions, or protected leave for layoffs announced earlier this year.

The lawsuit, reported by numerous media outlets in the United States, was filed in the US District Court for the Northern District of California in Oakland. It seeks a preliminary injunction to halt the layoffs before they take effect on July 22 while the employees pursue their claims through private arbitration.

The anonymous plaintiffs claim Meta relied on AI-generated productivity and performance assessments that penalised employees who had taken medical, parental, or family leave. 

They argue the process violated federal and state laws prohibiting discrimination and retaliation against workers with disabilities, pregnancy-related conditions, or legally protected leave. The lawsuit also alleges Meta failed to test its AI systems for bias, as required under recently adopted California and New York City regulations. 

Reports state that, according to the complaint, Meta used several internal AI-assisted tools to score and rank employees for potential termination. 

These allegedly included Metamate, the company’s internal large-language-model assistant; an employee-trained ‘second brain’ that indexed workers’ communications and documents; AI token usage metrics; and productivity scores generated by analysing keystrokes, screen activity, emails, and browser history. 

The plaintiffs argue those systems failed to account for periods when employees were absent due to protected leave, resulting in lower performance scores.

However, Meta rejected the allegations. Andy Stone, VP of Communication at the company, called the claims ‘untrue’. “Workforce management and organisational decisions were and are made by people, not AI,” he wrote on X.

The legal action follows Meta’s broad restructuring announced in May, when the company cut approximately 10% of its global workforce, affecting nearly 8,000 employees, while transferring about 7,000 workers into AI-focused teams as part of a wider reorganisation around AI products and internal AI-driven workflows.

Separately, Meta has also faced employee criticism over other internal AI initiatives. Earlier this year, more than 1,600 employees signed a petition opposing the company's Model Capability Initiative (MCI), which collected employee interaction data, including keystrokes, mouse movements, and screen activity, to help train AI models. Meta later paused the programme following a data security incident.