A significant shift in how Big Tech manages its workforce is unfolding in Ireland, as hundreds of workers tasked with “teaching” artificial intelligence face sudden unemployment. Documents obtained by WIRED reveal that Covalen, a Dublin-based firm providing essential services to Meta, is preparing to lay off over 700 employees.
This move highlights a growing trend in the tech industry: the use of human labor to build the very systems that will eventually render that same labor obsolete.
The Role of the “Human in the Loop”
The majority of the affected staff—roughly 500 individuals—work as data annotators. Their role is critical to the safety and functionality of Meta’s AI models. They perform “RLHF” (Reinforcement Learning from Human Feedback), a process where humans review AI-generated content to ensure it adheres to safety guidelines.
The nature of this work is both essential and psychologically taxing. Workers are tasked with:
– Testing guardrails: Creating complex prompts to see if an AI will bypass safety filters.
– Moderating extreme content: Reviewing material related to illegal acts, self-harm, and child exploitation to teach the model what to reject.
– Defining “correctness”: Acting as the gold standard for the AI to emulate.
One anonymous employee described the irony of the situation, stating, “It’s essentially training the AI to take over our jobs.”
A Strategic Pivot Toward Automation
While Covalen cited “reduced demand” as the reason for the cuts, the broader context suggests a massive strategic pivot by Meta. The company recently announced plans to nearly double its spending on AI technology, even as it implements sweeping layoffs across its wider operations to increase efficiency.
Meta’s strategy appears to be a transition from outsourced human oversight to internalized, automated systems. A spokesperson for Meta confirmed this direction, noting that the company aims to deploy “more advanced AI systems” to transform content enforcement, thereby reducing its reliance on third-party vendors like Covalen.
Impact on Workers and Labor Rights
The layoffs have been described by employees as abrupt and “undignified.” Reports indicate that workers were notified via a brief video meeting where they were not permitted to ask questions.
Beyond the loss of income, the Communications Workers’ Union (CWU) has raised concerns regarding the mobility of these workers. They report that affected employees face a six-month “cooldown period,” preventing them from applying to any other vendor that services Meta. This restriction significantly limits their ability to find immediate replacement work in a specialized field.
This is not an isolated incident for Covalen; this is the second major round of cuts in recent months. Following a strike in November, the company’s Dublin headcount is expected to be nearly halved.
Why This Matters
The situation at Covalen serves as a microcosm of the broader AI revolution. As companies race to develop more capable models, they rely heavily on a massive, often invisible, workforce of human annotators to ensure safety and accuracy. However, the ultimate goal of these companies is to reach a level of “artificial general intelligence” or highly sophisticated automation where human intervention is no longer required.
This creates a paradoxical cycle: human workers are being hired to build the tools that will eventually eliminate their roles.
As AI moves from a tool to an autonomous agent, the “human-in-the-loop” is being phased out in favor of more efficient, scalable, and cheaper automated systems.
Conclusion
The layoffs at Covalen signal a decisive shift in Meta’s operational model, moving away from human-led content moderation toward automated AI enforcement. This transition highlights the precarious nature of “training” jobs in the era of rapid automation.
