Twenty-six Meta employees say an AI-driven layoff process punished workers on medical and family leave, and a judge just let Meta keep cutting jobs while the case proceeds.
Story Snapshot
- Workers allege Meta used artificial intelligence to pick layoff targets in a biased way.
- A federal judge declined to block further layoffs during the lawsuit.
- Federal civil rights rules still apply when employers use AI, including in layoffs.
- The case shows how hard it is to prove whether AI actually made the call.
Employees Claim Biased AI Drove Layoffs At Meta
On July 14, current and former Meta employees filed a lawsuit claiming the company used artificial intelligence tools to help choose who to lay off. The 26 plaintiffs say the system targeted people who were on medical, pregnancy, parental, disability, or family caregiving leave. They argue the company’s use of automated scoring or ranking led to illegal bias against protected workers. Meta has not publicly conceded that any AI tool made the final decision in these cuts.
The plaintiffs say the tools were not properly tested and that Meta relied on automated assessments in ways that harmed disabled employees and those on protected leave. The complaint describes a process where internal systems weighed performance, availability, or other metrics that may penalize time away from work. The workers aim to show those tools, and how managers used them, caused a disparate impact on protected groups. That link is the crux of the case.
Judge Declines To Halt Layoffs As Case Moves Forward
On July 17, a federal judge rejected the workers’ request to stop Meta from continuing layoffs while the lawsuit proceeds. That ruling keeps pressure on the plaintiffs to gather evidence during discovery to prove whether and how artificial intelligence tools influenced who lost their jobs. The judge’s decision did not resolve the claims, but it means the company’s staffing moves go on while the legal fight continues.
Legal experts say the toughest task is showing causation. Workers must connect a data tool or automated score to a specific adverse action. In many cases, employers say artificial intelligence was only advisory, not decisive. That defense raises the bar for proof and drives requests for internal code, data, and emails. The Meta dispute spotlights this gap: public claims of bias versus the opaque systems that may sit behind a layoff choice.
Federal Rules: AI Use Does Not Erase Civil Rights Duties
The United States Equal Employment Opportunity Commission states that federal law protects workers when employers use artificial intelligence in layoffs or terminations in ways that discriminate on protected grounds, including disability. The agency’s worker guidance makes clear that automated systems do not excuse unlawful outcomes. Employers remain responsible for tools they buy or build and for how managers use their output in real decisions.
At an earlier public meeting, the Equal Employment Opportunity Commission heard testimony warning that artificial intelligence can codify and scale individual bias when trained on flawed data. That warning reflects a broader concern: if a model learns from past practices that undercount leave or accommodations, it can pass those penalties forward at speed. The hearing record frames why these disputes are rising as more employers rely on automated evaluation tools.
Why This Case Matters To Every Worker And Employer
This lawsuit fits a growing pattern. Employees challenge automated workplace decisions under long-standing civil rights laws. Employers counter that human leaders still decide. The Meta case is notable because it targets layoffs, not just hiring, and focuses on disability and protected leave. If the plaintiffs prove that automated scoring or rankings drove who got cut, it could shape how companies document and test any tool that touches jobs.
The road ahead runs through discovery. Plaintiffs will seek logs, model outputs, instructions to managers, and internal reviews. Meta will likely argue managers used multiple inputs and that no tool targeted protected leave or disability. For readers, one thing is settled: civil rights rules still apply, even when a machine helps make the call. The question now is factual, not philosophical—what the systems did in practice, and who was harmed by the results.
Sources:
americanbar.org, aclu.org, reedsmith.com, youtube.com, cnbc.com
