Shocking reshuffle within Meta

Internal emails are okay Business Insider Collected data shows that these 7,000 people were selected to join an initiative directly implemented by CEO Mark Zuckerberg, with the goal of promoting Meta’s position in the global AI race.

Many of the employees will be on the Applied AI (AAI) team, which Meta created earlier this year and is led by Vice President of Engineering Maher Saba, reporting directly to Chief Technology Officer Andrew Bosworth. Others were assigned to teams specifically focused on AI agents, such as the “Accelerating Agent Transformation” or “Data and Agent Optimization” teams.

“This reflects your influence,” the email said, explaining the recipient was chosen for his “high performance and technical ability, with the ability to make a difference to the new team.”

On May 20, about 8,000 people, equivalent to 10% of Meta’s workforce, began receiving layoff emails. They were told the cut was to “compensate for other investments” of the company. In that context, the transfer to the AI ​​department brings a ray of hope to many workers who are worried about losing their jobs.

Meta is implementing a series of large-scale reforms, such as establishing a super intelligence research department, organizing small teams dedicated to using AI, and changing the titles of some departments to “AI builders”. The reason is that Meta’s AI model is still behind competitors like OpenAI and Google, so company leaders have to find ways to mobilize human resources to train the model better.

Joining the new initiative is a “non-negotiable” move. Four anonymous employees said many people received the news in mixed moods such as sighing with relief, fear, confusion or shock.

“I was forced to hire,” one person wrote on a Discord channel dedicated to Meta employees discussing the layoffs. The Blind platform is also flooded with posts from those chosen to join AAI, some questioning what role this department really has.

Two Meta employees said these AI teams are likely in charge of labeling data, a job traditionally reserved for low-cost labor, which requires viewing each photo to annotate, or correcting errors in chatbot responses. This is an area in which Alexandr Wang, Meta’s Chief AI Officer, has a lot of experience. He is the co-founder and former CEO of Scale AI, one of the largest data labeling companies in the world, but they rely on a large number of subcontractors to train AI.

Meta spokesman declined to comment on the information.

An internal company announcement in April showed that AI models still lack the knowledge needed to surpass humans in many technical areas such as programming. “It is necessary to train models based on real-life examples so that they understand how people actually get work done,” the announcement reads.

In a leaked recording after an internal Meta meeting last month, Zuckerberg said Wang understood the business of labeling data very well. He also frankly commented that Meta employees “have much higher intelligence” than contractors, expressing his desire to mobilize top human resources across the company to train AI models.

“This will be a huge advantage if we can do it,” Zuckerberg said at the time.

Before that, according to Reuters, Meta also deploys many other AI initiatives, such as the Model Capability Initiative (MCI) tool that tracks employee mouse and keyboard movements. MCI works on work-related apps and websites. Besides collecting data, it also periodically takes screenshots of the content on employees’ screens. The goal is to improve the ability to simulate how humans interact with computers, such as selecting from drop-down menus or using keyboard shortcuts.

A series of Meta employees distributed leaflets to express their dissatisfaction. “You don’t want to work at the Employee Data Extraction Factory, do you?”, the leaflet stated and encouraged signing a petition opposing the installation of mouse and computer tracking software.

When asked about the above situation, Meta spokesperson Andy Stone said: “If we build agents that help people complete everyday tasks using computers, the models need real-life examples of how they use the machine such as moving the mouse, pressing buttons and manipulating drop-down menus.”

“Meta employees are paying the price for the reckless and costly gambles of the leadership team. While executives pursue a speculative AI strategy, employees face severe job cuts, harsh supervision and the cruel reality of having to train the ineffective system that is about to replace them,” Eleanor Payne, a member of United Tech and Allied Workers – a union founded by a group of Meta employees in the UK, told Reuters.