Reuters Reveals Meta Scrapped Ambitious AI Workforce Plan
Meta reportedly abandoned a plan to use AI agents to dramatically reduce team sizes after employee resistance, technical problems and weaker-than-expected productivity gains.
Meta CEO Mark Zuckerberg reportedly launched an internal initiative called “Project OT” in early 2026 to transform the company into a more AI-driven organization. According to Reuters, the plan envisioned autonomous AI agents handling routine work while smaller groups of employees supervised these systems. Some teams were potentially expected to shrink by as much as 60%, creating what Meta described internally as more “talent-dense” groups.
The ambitious strategy encountered significant resistance from Meta employees. Workers reportedly became concerned that they were being asked to help develop and train AI systems that could ultimately replace parts of their own jobs. Internal employee sentiment deteriorated, creating pressure on management to reconsider the speed and scale of the transformation.
Technical performance also became a problem. Meta discovered that its AI agents were not consistently producing the productivity improvements expected from the initiative. AI-generated code and automated workflows sometimes created additional complications, including excessive changes to software and security concerns. As a result, the company ultimately abandoned the planned second wave of workforce reductions scheduled for November, after an earlier 10% workforce reduction in May.
The reversal comes despite Meta continuing to make enormous investments in artificial intelligence. The company is spending heavily on data centers, computing infrastructure and AI talent, with Reuters reporting that its capital expenditure could approach $170 billion next year. Meta therefore remains deeply committed to AI even as it becomes more cautious about replacing human workers with autonomous systems.
The episode highlights an important lesson about the current AI revolution: building powerful AI is not the same as successfully integrating it into a large organization. AI agents may eventually automate significant amounts of knowledge work, but companies still have to deal with reliability, security, employee resistance and the economics of deployment. Meta’s experience suggests that the transition toward AI-native workplaces may be more complicated—and potentially slower—than the most aggressive automation predictions suggest.
