Meta’s plan to reduce staff with AI halted after employee backlash and productivity issues

Meta’s AI-centered restructuring plan has not proceeded as intended due to employee backlash and the technology's failure to deliver expected productivity. According to internal documents obtained by Reuters and sources familiar with the matter, company management evaluated scenarios for downsizing some teams by up to 60 percent through an initiative codenamed “Project OT.”

The plan reportedly envisioned shifting a significant portion of daily tasks at the parent company of Facebook and Instagram to AI tools and autonomous agents, with smaller human teams overseeing this virtual workforce. The restructuring was planned to be carried out in two waves, in May and November.

However, while Meta implemented a cut affecting 10 percent of its workforce on May 20, it shelved the second wave prepared for November. According to sources, this decision was influenced by internal unrest, negative indicators regarding the performance of AI tools, and investor questioning of rising costs.

COMPANY CONFIRMED THE EXISTENCE OF THE PLAN

In a statement to Reuters, Meta confirmed the existence of Project OT. The company stated that the effort was a year-long project focused on reducing costs, redesigning team structures, and shifting employees to priority areas such as generating training data for AI models.

The company also acknowledged that in the most severe scenarios, downsizing by up to 60 percent in some teams was discussed; however, it emphasized that this did not mean a target of laying off 60 percent of the entire workforce. Meta did not disclose which units were evaluated under these scenarios.

In its statement, Meta indicated that as part of the restructuring, some teams were asked to conduct scenario planning regarding task changes, the closure of open positions, and the potential impact of cuts. The company commented, “Ultimately, we did not implement all the scenarios in the exercise, and we did not have such an assumption anyway.”

Internal documents highlighted a working model with smaller “cell” teams supported by AI tools. In some pilots, teams consisting of two or three engineers and one designer were aimed at developing prototypes with shorter work cycles. By June, it was reported that at least 11 units, including engineering and research, had transitioned to this structure.

One of the issues that fueled employee backlash was software that tracked the keyboard and mouse movements of personnel in the US. Some employees expressed their reactions to management on internal communication channels, believing they were being tasked with training the AI systems that could replace them. It was noted that the internal satisfaction rate dropped from 74 percent to 55 percent.

Indicators regarding the productivity of AI tools also slowed the plan's pace. According to data shared by Meta Chief Technology Officer Andrew Bosworth, while code changes in the internal software infrastructure increased by 220 percent year-over-year, the increase in new or updated features reaching users remained at 36 percent.

According to sources, the uncontrolled increase in code caused by AI also raised reliability and security issues. It was reported that major technical and security incidents increased by 40 percent compared to the previous year, and the time employees spent resolving these issues increased by 70 percent.

It was stated that in a message Zuckerberg sent to staff following the May 20 cuts, he said he did not expect any new company-wide layoffs this year. Meta management also discontinued the application that tracked mouse movements, allowed some employees to return to their former teams, and moved to increase in-office social spending.

Nevertheless, cost pressure on the company continues. According to LSEG data, Meta's plan to invest at least 130 billion dollars in AI chips and infrastructure this year keeps investor questions about spending and cost-cutting measures alive.