Meta cancels AI workforce transition plan after internal disruption

Meta halted its AI-driven headcount reduction initiative following internal system instability that increased technical incidents and employee workload.

Meta canceled Project OT in May 2026 after initiating a first round of layoffs in May 2026. The initiative, launched in January 2026, aimed to reduce up to 60 percent of headcount through AI automation but was suspended before finalizing layoff numbers. Internal reports showed AI agents caused a 40 percent year-over-year increase in major technical and security incidents, while employee time spent resolving these issues rose by up to 70 percent. Meta also launched third-party AI agent sales in June 2026 after canceling the second round of layoffs. The company’s internal 'AI native' definition included automated workflows and AI-first feature development, with Andrew Bosworth reporting 220 percent year-over-year increases in internal code changes and 36 percent in user-facing feature changes.

This friction arises from Meta’s attempt to scale AI workforce replacement without resolving system instability. The 40 percent incident spike and 70 percent time increase indicate significant operational strain when deploying AI agents across complex human workflows. For social safety nets, this represents a potential barrier to smooth AI workforce transitions—meaning resources intended for human support systems could be diverted to manage AI-induced disruptions.

What to watch: Specific teams affected by layoffs, the exact reason for cancellation, and how Meta resolves the 40 percent incident increase. Caveats: Meta did not confirm which teams were impacted; the cancellation reason remains undetermined; Project OT’s full AI deployment scope was incomplete before cancellation; 220 percent and 36 percent metrics originated from internal posts by Andrew Bosworth.

Source: Ars Technica