Meta reportedly abandoned an AI-focused restructuring plan that would have laid off thousands
Meta reportedly paused or abandoned a proposed AI-led restructuring that could have cut thousands of jobs, after its internal AI tools failed to deliver the productivity gains expected and employees objected to workplace monitoring. The plan matters because it illustrates the practical limits and workforce tensions facing companies seeking to replace or reorganise human roles around AI agents.
Known as Project OT, the proposal envisaged smaller human teams supervising AI, broader “builder” roles and reductions of up to 60% in some teams; a further November lay-off round was reportedly considered. Meta said code changes to its AI platforms rose 220% year on year, but user-facing features increased only 36%, while technical and security incidents rose 40% and time spent resolving them grew 70%; the company has not ruled out future, smaller cuts.
- Meta reportedly halted major AI-driven job cuts after disappointing results.
- AI output gains were offset by more technical and security problems.
- Broader restructuring and future reductions may still be considered.
New here? Start with this
Meta is a technology company that owns Facebook, Instagram and WhatsApp, and employs people across areas such as software, advertising and safety. Like many large firms, it is investing in artificial intelligence tools that can help carry out tasks traditionally done by staff.
Project OT was a proposed change to how some teams worked, with more reliance on AI tools and employees taking on broader roles. Reports said the plan also raised concerns among staff about job losses and workplace monitoring.
The debate matters because companies are weighing possible efficiency gains against the reliability of AI systems and their effects on workers. Meta’s experience is part of a wider question about how quickly AI can take on workplace tasks and what human oversight is still needed.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Meta was correct to abandon this restructuring because the anticipated productivity gains failed to materialise proportionally—whilst code changes rose 220%, user-facing features increased only 36%, suggesting AI tools were not reliably converting internal productivity into user value. Employees' concerns about workplace monitoring reflect legitimate objections to pervasive surveillance and autonomy restrictions. Abandoning a restructuring predicated on unproven technology and facing workforce resistance was sensible.
The case against
Meta's decision to abandon the restructuring was excessively cautious, as the 220% increase in code changes demonstrates genuine AI productivity gains. Lags in user-facing features likely reflect factors beyond AI capability—testing cycles, deployment caution, and strategic prioritisation rather than tool failure. Companies that fail to restructure around substantial AI productivity improvements risk falling behind competitors with leaner, AI-enabled operations. Smaller teams with better oversight could actually address rising technical incidents and improve system reliability.