New York: Meta CEO Mark Zuckerberg entered 2026 with an ambitious idea that could have changed the way one of the world’s biggest technology companies works. The plan was to make Meta an artificial intelligence driven company where AI systems and virtual workers would handle much of the routine work now performed by thousands of employees.
The plan, known internally as Project OT, was developed after senior Meta executives explored how far the company could go in reshaping its workforce around AI. Some of the most aggressive scenarios considered reducing the size of individual teams by as much as 60 percent. The plan did not mean cutting 60 percent of Meta’s entire workforce, but it showed how seriously the company was considering a future with far fewer human workers.
The transformation was supposed to happen in two stages. The first round came in May, when Meta cut about 10 percent of its workforce, or roughly 8,000 jobs. A second restructuring was planned for later in the year.
That second stage, however, was eventually abandoned.
The decision came as Meta faced growing evidence that its AI push was not producing the dramatic productivity gains that executives had hoped for. Employees were using AI tools to generate far more computer code, but much of that additional output was not translating into a similar increase in new products and improvements for users.
Internal figures showed that changes to Meta’s software and infrastructure increased sharply, while changes that resulted in new or upgraded features for users grew at a much slower pace. The difference raised questions about whether AI was genuinely making employees more productive or simply allowing them to produce more work that still needed to be checked and corrected.
There were also growing concerns about reliability and security. As AI systems became more deeply involved in coding and other tasks, employees faced more technical problems and spent more time dealing with incidents. This created an uncomfortable situation for a company that had hoped AI would reduce the need for human intervention.
Meta’s workforce was also becoming increasingly uneasy about the changes.
Employees had already faced years of restructuring and job cuts. The prospect of AI taking over more of their work added another layer of uncertainty. Morale suffered as workers questioned whether the new technology was being introduced to help them or eventually replace them.
The company’s attempt to use employees themselves as training material created another controversy. Meta introduced a system that collected information about how some workers used their computers, including mouse movements, clicks and keystrokes. The purpose was to teach AI systems how people perform everyday computer tasks.
The initiative faced strong opposition from employees. Concerns increased after a security problem raised questions about who could access some of the information collected through the programme. Meta later paused the initiative while it investigated the issue and introduced additional controls.
The controversy highlighted a difficult problem for Meta. The company needs huge amounts of high quality data to build more capable AI systems, but collecting that data from its own workforce can create serious questions about privacy, trust and employee rights.
Despite these setbacks, Zuckerberg has not backed away from AI.
Meta continues to spend enormous amounts of money on data centres, computing equipment and AI research. The company has also invested heavily in recruiting some of the world’s leading AI researchers and engineers. Its Superintelligence Labs has become one of the most important parts of the company’s long term strategy.
Ironically, even Meta’s AI division is now being reshaped. The company recently moved to cut about 600 positions from its AI organisation as executives sought to make the group smaller and faster. The aim is to create teams in which individual employees have greater responsibility rather than simply building ever larger groups.
This suggests that Zuckerberg’s vision has not disappeared. It is changing.
The company appears to be moving away from the idea that AI can immediately replace huge numbers of workers and toward a model in which a smaller number of highly skilled employees use increasingly powerful AI systems to accomplish more.
For Meta, the lesson from Project OT may be that replacing people is more complicated than replacing tasks.
AI can write code, analyse information and perform repetitive work at extraordinary speed. But the technology still needs people to set priorities, check results, solve unexpected problems and take responsibility when things go wrong.
The experience could also serve as a warning to other companies racing to automate their workplaces. Producing more output does not necessarily mean creating more value. A company can generate more code, more documents and more automated actions while becoming less reliable and more difficult to manage.
Zuckerberg still believes AI will transform Meta. But the events of 2026 suggest that the transformation will not happen as quickly or as simply as some of his early plans imagined.
Meta is still betting billions of dollars on artificial intelligence. What has changed is the confidence that fewer employees automatically means a more productive company.
For now, the future at Meta looks less like a workplace without people and more like a workplace where people and AI are being forced to learn how to work together.