26 former and current Meta employees have filed a class-action lawsuit in the U.S. District Court for the Northern District of California, accusing the company of discriminatory use of artificial intelligence in selecting candidates for layoffs. The plaintiffs, acting anonymously due to fears of employer pressure, allege that algorithms disproportionately targeted individuals on maternity, family, or medical leave, as well as those who received special work accommodations due to disability.
The lawsuit was filed on July 13, and the plaintiffs received layoff notices on May 20 as part of a massive reduction of approximately 10% of Meta's workforce — about 8,000 people. Actual termination of employment for most is scheduled for July 22. According to the plaintiffs, Meta compiled the layoff list not based on weighted manager decisions, but using a combination of internal AI tools, including the corporate assistant Metamate, employee-trained "second brain" agents, data on AI token usage, automated performance ratings, and information on corporate device activity.
Leave as a Risk Factor
According to the lawsuit, the metrics used considered the volume of work completed, activity levels, number of code changes, and annual review results. Employees on medical, family, or maternity leave could not accumulate these indicators on par with working colleagues. The plaintiffs claim that Meta did not adjust assessments to account for legitimate periods of absence and did not conduct individual verification of results. One plaintiff — a scientist on approved leave before childbirth — received a layoff notice two days before giving birth. Another employee stated that their rating was lowered due to a break related to an injury, while a manager on medical leave was included in the layoff list on the 16th day of absence.
Meta's Position and Plaintiffs' Demands
A Meta representative called the claims unfounded, stating that "personnel management and organizational structure decisions were and are made by people, not AI." However, the plaintiffs cite laws prohibiting discrimination against pregnant individuals, people with disabilities, and employees using family or medical leave. They also emphasize that Meta did not audit the automated systems for bias. In the first stage, the plaintiffs demand a temporary halt to layoffs, preservation of salaries, health insurance, stock rights, and leave status until the proceedings conclude, as well as an independent audit of the AI tools used. Their lawyers warn that once layoffs are completed, the consequences will become irreversible: employees will lose health insurance during pregnancy, postpartum recovery, and treatment, unvested stock rights, and, in some cases, immigration status.
Context: Data Collection for AI Training
In April 2026, Meta began installing the Model Capability Initiative system on corporate devices, which could record keystrokes, mouse movements, screen content, browser history, messages, emails, and for some employees, audio, video, and location data. Initially, the company stated that this data would only be used for training AI agents, not for performance evaluation. However, after widespread discontent from over 1,600 employees who signed a petition alleging privacy violations, Mark Zuckerberg suspended the program in June. The plaintiffs claim that the company announced the program through a low-profile internal post, and in some departments, employees were not asked for consent and were initially not given the option to opt out.
My expert opinion: This lawsuit is just the tip of the iceberg. Tech giants are increasingly using AI to optimize HR decisions, but the transparency and fairness of such systems leave much to be desired. If the court finds Meta's actions discriminatory, it will set a precedent that forces the entire industry to reconsider approaches to automating personnel management. Companies must remember: algorithms are a tool, not a replacement for human judgment, especially when people's lives and careers are at stake.