Meta's Applied AI Team Faces Internal Revolt Over Work Conditions

Tensions Boil Over in Company Meeting

An employee interrupted a livestreamed Meta presentation this week with an expletive-filled outburst, calling out leadership and declaring frustration with "being the company's bitch," according to a recording obtained by WIRED. The individual directed presenters to tell a specific AI executive "that he's a piece of shit" before the meeting continued.

The incident reflects mounting discontent within Meta's Applied AI team, a 6,500-person unit formed in March to support AI research at Meta Superintelligence Labs. Three current employees told WIRED that widespread dissatisfaction stems from how the company assembled the group and the nature of work assigned to them.

Engineers Describe 'Soul-Crushing' Assignments

Members of Applied AI report being tasked with generating puzzles and coding problems to test and train AI models—work they characterize as menial compared to their previous software development roles. "It's literally the gulag," one employee said. "You have zero purpose in life all of a sudden, you barely interact with anyone, you just have these tasks every week."

Another employee described the assignments as easy but unfulfilling. "Most people find the work soul-crushing," a third team member said, adding that "almost all" employees appear unhappy.

Unlike typical Silicon Valley practices, engineers selected for Applied AI must either join the unit or leave the company entirely. Some members refer to themselves as "draftees" due to this lack of choice. The team has grown in waves since early April, with some workers required to complete two tasks per week.

Company-Wide Morale Issues

The Applied AI situation is part of broader organizational turmoil at Meta. The company's AI-focused restructuring included eliminating 10 percent of the workforce—8,000 employees—last month. More than 1,600 employees have signed a petition opposing a new initiative that monitors US workers' clicks and keystrokes to generate AI training data.

During an Instagram all-hands meeting this week, Meta chief product officer Chris Cox acknowledged the "difficult" and "brutal" environment, comparing recent months to "running a marathon in the middle of a hailstorm." Cox said leadership needed to "get in touch with the company again" and avoid being "overearnest" about AI's capabilities.

Zuckerberg Responds

In an internal memo Friday, CEO Mark Zuckerberg acknowledged that organizational changes had caused distress and admitted the company had "made mistakes." He pledged no additional mass layoffs this year and announced plans to reduce manager-to-employee ratios, which had reached 50-to-one on some teams including Applied AI.

Zuckerberg characterized Applied AI work as "critical to advancing our models" and suggested the team serves as a temporary assignment while the company creates other roles. He outlined Meta's AI vision as focused on personalized social media experiences, smart glasses, small business tools, and "personal superintelligence agents."

Why it matters

Meta's internal turmoil reveals the human cost of rapid AI pivots at major tech companies. As organizations race to compete in generative AI, the reassignment of thousands of skilled engineers to data labeling and model evaluation work raises questions about talent retention and whether such aggressive restructuring can sustain innovation. The company's approach—forcing engineers to accept new roles or resign—represents an unusually rigid strategy that could influence how other tech firms manage AI transitions.

These details were first reported by WIRED.