Algorithms don’t have opinions. They don’t hold grudges. They also don’t know — or care — that you spent three months on medical leave for a disability or a complicated pregnancy. Twenty-six former Meta employees now allege that’s exactly the problem. Filed in federal court in Oakland, their lawsuit claims Meta used AI-powered productivity metrics, reportedly including measures tied to AI usage and output, to select workers for cuts during the company’s 2026 layoffs. The result, according to the complaint: a system that looked objective but allegedly filtered out people with medical histories like a bouncer working from a list no one publicly posted.
The Algorithm in the Room
Anonymous plaintiffs from six states say Meta’s AI-linked scoring quietly penalized disability and medical leave.
The plaintiffs filed anonymously, spanning California, New York, the District of Columbia, and three other states. Their central claim is straightforward: Meta’s productivity scoring allegedly didn’t account for the fact that workers on medical leave or disability accommodation naturally generate lower output metrics. According to the complaint, the system treated absence like underperformance.
- Twenty-six plaintiffs filed in U.S. federal court, Oakland, California
- Plaintiffs come from six states, including California and New York, plus the District of Columbia
- Allegations span violations of federal and state disability, medical leave, and pregnancy-discrimination laws
- Meta announced plans to cut roughly 8,000 jobs — about 10 percent of its global workforce — beginning in May 2026
- The EEOC has stated that existing anti-discrimination law applies when employers deploy AI in workforce decisions
Meta’s spokesperson told Reuters the claims lack merit, saying workforce decisions “were made by people, not AI.” That defense sounds clean until you consider what analysis from the Cornell Journal of Law and Public Policy flagged in 2024: AI tools in HR can reproduce and amplify human bias rather than eliminate it. A human approving an algorithm’s output doesn’t automatically cleanse what that output was built to measure. If the scoring criteria invisibly penalize time away from a keyboard, the human signature at the bottom doesn’t change what happened upstream.
Pattern Recognition
This isn’t Meta’s first discrimination lawsuit tied to its 2026 layoffs.
A separate suit reported by the Mercury News in March 2026 alleged Meta’s cuts disproportionately affected older workers. Courts are equipped to recognize patterns across suits the same way regulators trace systemic bias — methodically, and with documentation. If this case survives early dismissal motions, it could pressure large employers across the industry to audit their AI tools and wall off protected medical data from productivity analytics entirely.
The courtroom question is whether Meta broke the law — courts will decide that. The broader question worth sitting with is whether your employer’s performance dashboard has the same blind spots baked in, and whether anyone thought to check before the next round of cuts.
