Metaのレイオフ判断はAIが行ったと訴訟が主張Lawsuit claims Meta's layoff decisions were made by AI, not humans

A newly reported lawsuit against Meta alleges that the company leaned on artificial intelligence to decide which employees to lay off, with the plaintiffs claiming those decisions were made without meaningful human review. Covered by Ars Technica, the case could become an early test of how existing employment law applies when automated systems shape or determine who keeps a job, a question that has moved quickly from theoretical debate to courtroom dispute.

According to the complaint, former Meta workers contend that internal systems used to rank, score, or evaluate performance fed directly into termination decisions, and that the process lacked the kind of oversight that would let affected employees understand or challenge the outcome. The plaintiffs appear to be arguing that when a consequential decision such as dismissal is effectively delegated to software, an employer cannot escape responsibility for errors, bias, or unfairness embedded in that software. Meta has not publicly conceded that AI made the final call on any individual, and the allegations remain unproven claims that the company is expected to contest.

The dispute lands against the backdrop of Meta's own recent history. The company cut tens of thousands of roles across 2022 and 2023 during a period Mark Zuckerberg branded the "year of efficiency," and it has continued to trim what it describes as lower-performing staff while investing heavily in AI infrastructure and talent. That combination, aggressive workforce management alongside a deep push into automation, is central to why this particular suit is drawing attention. Critics have long warned that companies embracing AI internally might use it not only to build products but also to manage, evaluate, and remove people.

Technically, the case touches on what are often called automated employment decision tools. These systems can aggregate performance metrics, productivity signals, peer and manager feedback, and other data to produce rankings or risk scores. The concern raised by researchers and regulators is that such tools can reproduce or amplify bias present in their training data, and that their internal logic is frequently opaque even to the organizations deploying them. The phrase "meaningful human oversight" is doing heavy lifting here, because in many workflows a manager may nominally approve a list that an algorithm generated, blurring the line between a human decision informed by data and a machine decision merely rubber-stamped by a human.

The legal questions are not entirely new, but they remain unsettled. In the United States, the Equal Employment Opportunity Commission has issued guidance warning that anti-discrimination laws still apply when AI tools are used in employment decisions, and that employers can be liable for discriminatory outcomes regardless of whether a vendor built the software. A closely watched case, Mobley v. Workday, has advanced the theory that an AI screening provider could itself face liability, signaling that courts are willing to examine how these systems function. New York City's Local Law 144 requires bias audits for certain automated hiring tools, and the European Union's AI Act classifies many employment-related AI systems as high-risk, imposing transparency and human-oversight obligations. The Meta suit, if it proceeds, is likely to draw on this evolving framework as plaintiffs seek to establish accountability.

For the broader industry, the significance lies less in any single company and more in the precedent such litigation could set. As firms across technology, finance, and retail experiment with AI for scheduling, performance management, and workforce planning, the boundary between decision support and decision making is becoming harder to define. A ruling that clarifies when an employer must provide human review, disclose the role of automated tools, or explain an adverse decision would ripple well beyond Meta.

At this stage, key facts remain contested, and the outcome will depend on evidence about how Meta's systems actually operated and how much discretion human managers retained. What the filing does illustrate is a growing willingness by workers and their attorneys to scrutinize the internal use of AI, not just its consumer-facing applications. Whether or not the specific allegations hold up, the case reflects a wider reckoning over transparency, fairness, and legal responsibility as automated systems take on a larger role in decisions that affect people's livelihoods.