Meta drafts 7,000 employees into Applied AI push

Meta drafts – As 8,000 Meta employees read layoff notices this week, about 7,000 others received an email inviting them to join Mark Zuckerberg’s new AI initiative, including the Applied AI team led by Maher Saba and agent-focused groups headed by Andrew Bosworth. The moves

On Wednesday, the same inboxes carried two very different messages inside Meta.

For 8,000 employees, it was the paper cut of layoffs: they were reading their layoff notices.

For 7,000 others, it was an email with a different kind of shock. Employees were told they’d been selected to join a new AI initiative spun up directly by CEO Mark Zuckerberg—part of a push to speed Meta’s race toward stronger AI systems.

Those reassignments landed while Meta’s workforce, which numbers 78,000, was being cut. For some employees, the chance to move into AI teams offered a lifeline. For others, the relief came tangled with uncertainty about what their new roles would actually look like.

The teams employees were being asked to join were not generic innovation projects. Many were moved into a group called Applied AI (AAI), created earlier this year. That group is led by engineering vice president Maher Saba and reports to chief technology officer Andrew Bosworth.

Other employees were pulled into work more specifically focused on AI agents. These include a group named “Agent Transformation Accelerator,” headed by Bosworth, and a team named “Agent Data and Optimization,” according to internal messages shared with Business Insider.

In at least one email copy reviewed by Business Insider, the message was framed as a recognition of past performance. “This is a reflection of your impact,” the email told recipients, adding that they were picked because of their “strong performance” and technical abilities.

“You were identified as someone who can make a real impact on this team,” the email said.

Internally, the effect of the week’s cuts and the draft into AI work showed up fast. In a Discord server where Meta employees discussed the layoffs, one person said, “I got drafted.” Another replied, “Welcome to the draft.”

The relief-versus-dread split was also visible in employee forums. On Blind, an anonymous workplace app, a section locked to verified Meta employees was flooded with posts from people announcing they had been drafted. Some asked what AAI actually does.

Meta did not provide a public response to clarify the initiative. A Meta spokesperson declined to comment.

What might happen inside the new teams

Behind the emails sits a practical question for employees who’ve been reassigned: what work will they do day to day?

Two Meta employees told Business Insider they expect the new AI task forces to involve data labeling—manual work such as tagging images or correcting chatbot responses.

That’s a domain where Meta’s AI chief, Alexandr Wang, has deep experience. Wang is the cofounder and former CEO of Scale AI, one of the world’s biggest data-labeling companies that relies on large numbers of contractors to train AI.

Meta has also been building additional internal tools and initiatives for AI work.

Business Insider previously reported that Meta launched an internal tool to track employees’ keystrokes and mouse movements, known as Model Capability Initiative.

An internal announcement in April about the software said that AI models still lacked the knowledge to outperform humans at technical tasks such as coding. That April posting also made the case for training on real behavior: “For agents to understand how people actually complete everyday tasks using computers. we need to train our models on real examples.”.

In a leaked audio recording from an internal meeting that same month. Zuckerberg described why he believed Meta could benefit from drawing on its own talent for AI training. Zuckerberg said Wang knew the data-labeling world well. then added a blunt comparison: the average Meta employee has a “significantly higher” intelligence than those contractors. He said he’d rather “enlist” top employees from across Meta to train its AI instead.

“I think that this is going to be a very big advantage, if we can do it,” Zuckerberg said, adding that it remained a “hypothesis” for the time being.

Now, the week’s draft letters suggest Meta is testing that hypothesis with real people—at a moment when every job change carries amplified stakes inside the company’s 78,000-person workforce.

Meta Zuckerberg Applied AI AAI Maher Saba Andrew Bosworth AI agents data labeling Alexandr Wang Scale AI Model Capability Initiative employee layoffs workforce reorganization