The spreadsheet made it to court. Meta is facing a class-action lawsuit, first reported by the Los Angeles Times, alleging that the company used an AI system to analyze employee attendance and leave records during layoffs, systematically singling out workers who were on family and medical leave. The employees who filed the suit say the system treated “leave status” as a factor that made workers more likely to be cut.
The details of the allegation are specific. According to the lawsuit, Meta’s AI screening model reviewed attendance and absence data and identified employees on protected leave as a priority category for layoffs, leading to a disproportionately high rate of job cuts among workers taking family and medical leave under the Family and Medical Leave Act, the U.S. law that guarantees workers time off for medical and family reasons. HR Executive, which covered the case, reported that the model flagged leave status as a layoff-priority feature.
The case joins a growing line of litigation over the use of algorithms in employment decisions. Companies have used automated tools for years to screen resumes, schedule shifts and evaluate performance, but the Meta suit targets a different use: an AI system applied to deciding who keeps their job. Federal law has long protected workers from retaliation for taking leave, and plaintiffs’ lawyers have argued that an algorithm that factors leave into layoff decisions is discrimination by another name.
Meta has not publicly responded to the lawsuit’s specific allegations. The company has previously said its layoff decisions are made by managers with attention to legal obligations, and it has defended its use of data-driven processes. The case is at an early stage, and the court has not ruled on whether the class should be certified.
The lawsuit lands at a delicate moment for Meta’s reputation on AI. The same week, the company’s Oversight Board, the independent body that reviews Meta’s content decisions, published a study concluding that leading AI models, including GPT-5.6, Claude and Gemini, visibly avoid critical statements when asked about repressive regimes. The report was widely read as a criticism of the industry’s willingness to self-censor, and Meta’s own AI products were included in the findings.
Taken together, the two stories paint a picture of a company whose AI systems are under scrutiny from both directions: the algorithm that decides who gets fired and the models that decide what can be said. Meta has made AI central to its strategy, investing billions in its own models and in the data centers that run them, and its executives have argued that AI will make its products and its operations more efficient. The lawsuit tests that argument in the most personal way possible.
The legal questions are far from settled. Employers are allowed to use data in layoff decisions, and the lawsuit will have to prove that Meta’s system did more than reflect decisions already made by managers. But the use of an algorithm that flags leave status creates a paper trail that a jury could read as intent, and employment lawyers said the case could reshape how companies design layoff algorithms even if Meta wins.
The context includes Meta’s history of large workforce reductions. The company cut more than 20,000 jobs across two rounds of layoffs in 2022 and 2023, and it has conducted smaller cuts since, including rounds tied to its AI reorganization. Each cut has produced litigation, and Meta has settled some claims and fought others. The new suit adds an AI dimension to a pattern the company knows well.
The plaintiffs’ argument has a simple logic. If a worker is on protected leave, they are not in the office, not visible in the daily workflow, and easier for a data-driven process to overlook. An AI system trained on attendance patterns can encode that bias without anyone writing a policy that says “fire people on leave.” The lawsuit’s central claim is that Meta’s system did exactly that, and that the company’s managers either knew or should have known.
The case also raises questions about how layoffs are conducted across the technology industry. Many large tech companies have described using data and analytics to guide workforce reductions, and the Meta suit is being watched by employment lawyers who represent workers at other firms. If the class is certified and the evidence supports the plaintiffs, the decision could affect how every company that uses HR data to plan cuts designs its systems.
For Meta, the timing compounds the problem. The company is fighting multiple fronts on AI: regulators in Europe examining its content moderation, advertisers pressing for changes to its targeting tools, and now employees arguing that its internal use of AI crossed a legal line. The suit adds a domestic legal exposure to a list that is already long.
The outcome will turn on details that are not yet public: what the AI system actually did, how it was built, who approved its use and what the layoff criteria were. The plaintiffs say the documents will show a system designed to deprioritize workers on leave. Meta will argue that the model was a tool, and that the decisions were human.
Whatever the court decides, the case has already changed the conversation. AI systems that decide who keeps a job, and that weigh factors like leave status without explicit human sign-off, are now the subject of a lawsuit that could define the rules of the road. For the thousands of tech workers who have been laid off by algorithm-driven processes in recent years, that is a question with a personal answer.
