Meta Tracks Employee Typing to Train AI, While Cutting Jobs to Fund New Technology

Meta has started monitoring what employees type, how they click the mouse, and how they move the cursor on company computers. The company is using this data to train artificial intelligence systems that could eventually do routine office work. At the same time, CEO Mark Zuckerberg has directly stated that Meta is laying off workers partly because the company is spending billions on AI infrastructure instead.

Zuckerberg made this connection clear during a company meeting on April 30, saying that as Meta invests heavily in AI, it needs fewer people doing certain jobs. When asked if more layoffs could come, he did not rule it out, suggesting that the company may continue cutting staff as it shifts money toward AI development.

How Meta Is Using Employee Data for AI Training

Meta is capturing detailed records of how employees work: the patterns of their keystrokes, how fast they type, where they move the mouse, which applications they switch between, and how long tasks take them. Think of it like recording a video of someone's hands while they work, then teaching a robot to copy those movements.

This data helps train AI agents — software programs designed to do routine computer tasks automatically. By learning from real employees doing real work, these AI systems can learn to navigate the kinds of messy, complicated software environments that people actually use in offices, rather than learning from simplified test scenarios.

The monitoring happens across Meta's computers and internal programs. The company is collecting information about how workers move through their daily tasks.

Why Meta Is Cutting Jobs to Spend on AI

Technology companies face a choice about how to spend their money. They can hire more people, or they can invest in computers and artificial intelligence that might do some of the work people currently do. Meta is choosing the second path.

Building the infrastructure for large-scale AI is expensive. It requires specialized computer chips called GPUs, large data centers, networking equipment, and highly skilled engineers to build and maintain it all. That costs billions of dollars per year.

Zuckerberg has framed this as a strategic decision: rather than hire more people, Meta is betting that AI will eventually be powerful enough to do more work with fewer employees. The company is trading payroll costs for technology costs.

The broader context here is that Meta and other major technology companies are making a calculation: invest heavily in AI infrastructure now, reduce the workforce, and hope that the AI systems become good enough at real work to justify the money spent and the jobs lost. Whether this bet will pay off remains to be seen.

This Pattern Has Happened Before

Technology companies have a history of looking inward at their own operations to find advantages. When the web became commercial in the early 2000s, companies studied how people clicked around websites and used that data to build better products. When smartphones took off, companies studied mobile usage patterns. Now, with AI, Meta is studying how employees actually work.

Each time, a company mines its own real-world data to improve its technology. In this case, Meta has access to something valuable: thousands of detailed records of how skilled workers actually use computers during their jobs.

The Competitive Advantage

If Meta can train AI systems using real employee behavior, those systems might be better at actually doing work than AI trained on test data or invented scenarios. By learning from genuine workplace patterns, the AI could better handle unexpected situations and the complicated reality of real office software.

This approach also gives Meta a potential edge against competitors who might be training their AI differently. If Meta's AI agents end up being more capable and more practical, the company could use them to automate more work and sell them as products to other businesses.

What Happens Next

The real test will be whether Meta's AI agents can actually do meaningful work reliably. If they can, the company's strategy of cutting jobs while investing in AI will look like a smart decision. If the AI turns out to be less capable than hoped, the company will have reduced its workforce without the productivity gain it was banking on.

The broader industry will be watching. Other major technology companies may follow Meta's example as AI capabilities improve, creating a template where workforce reductions are tied directly to AI infrastructure spending.