Meta employees sue over alleged AI bias in May layoffs

A group of 26 unnamed current and former Meta employees sued the company in federal court in California, alleging that artificial intelligence-linked systems used in a May workforce reduction discriminated against some workers. The plaintiffs say they were among the 10% of Meta employees cut in that layoff round, according to a complaint filed Monday in the U.S. District Court for the Northern District of California.

The complaint alleges Meta violated protected-leave and anti-discrimination laws tied to pregnancy, disability and other categories. The workers are seeking to pursue their claims individually in arbitration, while asking the court for interim relief that would preserve their employment status.

Meta rejected the allegations. A company spokesperson told CNBC by email that the “claims lack merit and are not based on facts,” adding: “Workforce management and organizational decisions were and are made by people, not AI.”

How the employees describe Meta’s layoff process

Lawyers for the plaintiffs allege that Meta relied on what they called a “constellation of internal artificial-intelligence systems” when selecting employees for termination. According to the complaint, those systems drew on data including performance reviews, calibration scores, productivity measures, output metrics, “AI-native” ratings and AI-token consumption.

Token consumption has become a measure of how much a worker uses AI tools. In the plaintiffs’ account, that kind of metric can penalize employees who are not actively working because they are on approved medical or family leave, or whose output is affected by a disability.

The complaint says the systems did not properly account for approved absences when measuring employees against those inputs. The plaintiffs allege that workers on leave could not generate the same volume of productivity signals, AI-use indicators or output data as employees who were continuously present, making them more vulnerable in the reduction process.

The filing asks the court to issue a preliminary injunction “maintaining the status quo of their employment” at Meta while an independent audit of the algorithmically assisted selection process is conducted and the arbitration claims are resolved.

Legal scrutiny of workplace AI grows

The Meta case adds to rising legal attention on employers’ use of AI tools in personnel decisions, particularly where automated or algorithmic systems rely on historical employment data or proxy measures that may reflect protected characteristics or protected absences.

Courthouse News Service previously reported the lawsuit.

The filing comes about a month after a federal judge in California ruled that Workday must face claims in a separate case involving AI-powered job screening tools. In that matter, plaintiffs alleged that Workday’s recruiting technology violated state and federal anti-discrimination laws.

Workday denied those allegations. The company said at the time that its AI recruiting software does not make hiring decisions “in California or anywhere else.” Workday also said its technology considers job qualifications rather than protected traits such as race, age or disability, and that it tests products under its Responsible AI program to check whether tools harm protected groups.

The Meta plaintiffs’ claims remain allegations at this stage. The court has not resolved the merits of their discrimination claims, and Meta says people, rather than AI, made its workforce management decisions.

This story draws on original reporting from CNBC.