Meta’s ambitious plan to replace human roles with AI-driven automation has collapsed after triggering mass layoffs, technical disruptions and a sharp decline in employee sentiment. The initiative, codenamed Project OT, aimed to restructure the company toward an AI-first model where virtual agents handled routine tasks under smaller human oversight teams. Internal documents and interviews indicate the plan backfired, leading to a 10% workforce reduction in May and widespread operational instability.
The restructuring followed Meta’s 25% staff cut between late 2022 and early 2023, but the latest overhaul escalated discontent. Executives explored scenarios involving up to a 60% reduction in some teams, though only a 10% layoff was executed on May 20. By the end of May, engineering units had seen a 30% drop in headcount. Internal data showed a 220% year-over-year surge in AI-driven code changes, contributing to a 40% increase in severe technical incidents, including outages and data leaks linked to unsupervised AI agents.
Employee morale deteriorated rapidly. Internal pulse scores fell from 74% to 55%, with workers reporting a 70% rise in time spent addressing AI-related disruptions. Meta’s infrastructure teams flagged "reliability warning signs" from sudden AI coding surges, while hackers exploited a new AI customer service bot to access high-profile Instagram accounts, including an inactive Obama White House page. The company later suspended a tool requiring U.S. employees to install keystroke-tracking software on their devices to train AI agents.
Zuckerberg acknowledged missteps in an internal meeting in early July, canceling planned November cuts. The retreat, held at his Hawaii property in January, had outlined a shift toward AI-native workflows, including the elimination of traditional roles like product designers and engineers in favor of a generic "builder" title. Meta now plans to invest $130 billion in AI infrastructure this year, but the failed restructuring has underscored the challenges of balancing automation with operational stability.
Internal communications revealed widespread frustration, with employees posting critical comments on Meta’s workplace platform and using the elephant emoji to symbolize the "elephant in the room." Chief Data Officer Alex Schultz and other executives have since emphasized a more measured approach to AI integration, though the episode has left lasting scars on the company’s workforce and reputation.
