Meta is laying off hundreds of employees responsible for improving AI... to replace them with AI

A video conference layoff 

Layoff waves continue at Meta. We learn today that 700 people working for the company Covalen have been laid off. This company, based in Ireland, handles content moderation and labeling for Meta. Among the 700 laid off, 500 are annotators. According to the Communication Workers' Union (CWU), Covalen's workforce will be reduced by nearly half when considering all layoff plans.

A video conference layoff

Moderators and annotators replaced by AI

To justify this layoff plan to the affected employees, Meta does not mention artificial intelligence and only refers to a « demand decline and operational needs ». In contrast, Erica Sackin, a Meta spokesperson, is more direct in a statement: « As we indicated in March, Meta will deploy more advanced AI systems over the next few years to transform our approach to content moderation and management across all our platforms, and thus ensure the security and protection expected by users. In doing so, we will reduce our reliance on third-party vendors and strengthen our internal systems ».

As mentioned previously, most of the laid-off employees are annotators. Their work involves, among other things, pretending to be a malicious user, with the aim of bypassing the safeguards imposed on the AI model. Nick Bennett, an employee at Covalen, explains «it’s a draining job ». In fact, « we spend our days pretending to be a pedophile ». 

The work of most of the laid-off employees was therefore to improve the AI models of Meta. An employee who wished to remain anonymous arrives at a cynical conclusion: « In reality, we are training the AI to take our place ». 

Moderation primarily operated by AI

One might wonder if the moderation primarily operated by artificial intelligence is relevant across different Meta platforms, especially when it comes to children under 13. To recall, Mark Zuckerberg's company would have generated billions of profits through fraudulent advertisements. Some will say that Meta deliberately allows these fraudulent ads to proliferate on its platforms for financial purposes, but the American company vehemently denies it.

👉🏻 Follow real-time tech news: add 01net to your Google sources, and subscribe to our WhatsApp channel.

Source : Wired