Meta's Applied AI unit, a 6,500-person team formed to support AI research, is rife with discontent as engineers report being assigned menial data-generation work rather than meaningful projects.

What happened: Meta assembled the Applied AI unit in March with about 6,500 engineers and product managers to support AI researchers. Three current employees tell WIRED the team is experiencing widespread dissatisfaction, with tasks described as generating puzzles to test AI models—work employees characterize as menial compared to their previous software development roles. An expletive-filled outburst during an employee livestream and more than 1,600 employees signing a petition over a separate employee-monitoring initiative reflect record-low morale across the company following a restructuring that included 8,000 layoffs last month.

Why it matters: Retaining top technical talent in AI is crucial for Meta's ability to compete in the growing AI services market. When highly skilled engineers feel their work is soul-crushing and uncreative, the company risks losing people it needs for its AI ambitions. CEO Mark Zuckerberg acknowledged in a memo that recent organizational changes had caused distress and that he needs to provide stability and give employees a sense of purpose to achieve Meta's goal of being "the best place for the most talented people in the world to make an impact."

What to watch: Zuckerberg plans to limit the number of employees per manager (Applied AI had a 50-to-one ratio on some teams), increase team event budgets, hold a large hackathon next month, and assign desks to employees in many locations by year-end. He also pledged not to carry out additional mass layoffs this year. The test will be whether these measures and clarification that Applied AI is a waypoint rather than a permanent role can stabilize morale and retain the talent the company has forcibly assigned to the unit.