‘It’s Disgraceful’: Hundreds of Employees Training Meta’s AI Face Potential Layoffs

Numerous employees in Ireland responsible for enhancing Meta’s AI models have been alerted that their positions are at risk as the company initiates a significant new phase of layoffs, based on documents acquired by WIRED.

The workers impacted are employed by the Dublin-based company Covalen, which manages various content moderation and labeling services for Meta.

They were notified of the layoffs during a brief video call on Monday afternoon, with no opportunity for questions, as reported by Nick Bennett, one of the participants on the call. “We had a pretty bad feeling [before the meeting],” he states. “This isn’t the first time it’s happened.”

In total, over 700 employees may face job losses at Covalen, according to an email seen by WIRED. Approximately 500 of these are data annotators responsible for evaluating content generated by Meta’s AI models against the company’s guidelines prohibiting harmful and illegal material. “It’s effectively training the AI to take our jobs,” asserts another Covalen employee, who requested anonymity out of concern for backlash. “We make decisions that the AI is designed to emulate.”

At times, the role also involves creating intricate prompts aimed at circumventing safeguards that prevent models from suggesting child sexual abuse material or suicide descriptions. “It’s quite a taxing job,” says Bennett. “You spend your entire day pretending to be a pedophile.”

Last week, Meta revealed intentions to eliminate one in ten jobs as part of extensive layoffs aimed at enhancing company efficiency. A memo circulated within the company reportedly indicated that these layoffs were driven by a need to increase investments in other business areas. While the memo did not mention AI, the company recently announced plans to substantially increase its spending on technology. In January, Meta CEO Mark Zuckerberg remarked, “I believe that 2026 will be the year AI begins to fundamentally alter the way we work.” The email reviewed by WIRED indicated that Covalen employees were informed that the layoffs stemmed from “reduced demand and operational requirements.”

In a statement, Meta spokesperson Erica Sackin remarked: “As we mentioned in March, over the coming years, Meta will be implementing more advanced AI systems to transform our strategy for content enforcement and operations across our platforms, ensuring it offers the safety and protection that people expect. As we move forward, we’ll be decreasing our dependence on third-party vendors and bolstering our internal systems.”

This round of layoffs marks the second instance in which Covalen has reduced staff in recent months. In November, the company announced plans for job reductions (reportedly around 400), which led to a worker strike. Between the two rounds of layoffs, Covalen’s workforce in Dublin is projected to be nearly halved, according to the Communications Workers’ Union (CWU), which represents some Covalen employees.

For the affected Covalen staff, the effort to find new employment will be complicated by a six-month “cooldown period,” during which they cannot apply to another Meta vendor, according to the CWU. “It’s undignified, you know,” says the Covalen employee who preferred to stay anonymous. “It’s rude.”

Covalen did not respond immediately to a request for comment.

Unions representing the affected workers are advocating for Covalen to engage in discussions about severance terms. They also aspire to meet with the Irish government to talk about the effects of AI on the workforce in the country. “Tech companies are treating the employees whose labor and data contributed to building AI as expendable,” states Christy Hoffman, general secretary of UNI Global Union. “To counter this, it’s absolutely essential that workers organize and demand transparency regarding the implementation of AI, training linked to employment, and a plan for their futures. Workers should also possess the right to refuse training their AI replacements.”

However, some impacted by the layoffs are skeptical about their prospects for finding stable jobs in a labor market continuously reshaped by AI and the financially robust companies driving its advancement. “It’s a fundamental struggle between struggling white-collar workers and large capital, really,” asserts Bennett. “That usually tends to favor the latter.”

Update 4/28/25 3:30pm ET: This article has been updated to incorporate comments from Meta.