Meta Is Watching Every Click: What the Tech Giant's New AI Training Tool Means for the Future of Work

The move, first reported by Reuters and confirmed to the BBC by a Meta spokesman, involves a new internal tool called the Model Capability Initiative (MCI). The tool runs on Meta's computers and internal applications, recording how workers navigate their daily tasks to generate real-world training data for AI agents being developed inside the company.

"If we're building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them," a Meta spokesman told the BBC.

A Workforce on Edge

The announcement has landed at a particularly sensitive moment inside Meta. The company has already eliminated roughly 2,000 positions in 2026, with employees bracing for a deeper wave of cuts in the months ahead. Meta's job listings page, which hosted approximately 800 openings as recently as March, now advertises just seven roles — a stark indicator of a sweeping hiring freeze.

Against this backdrop, the surveillance tool has drawn sharp reactions from inside the company. One current Meta employee, speaking anonymously to the BBC, described the experience as "very dystopian," adding that having their smallest computer actions logged while job cuts loom only heightened the sense of unease. A recently departed employee was more blunt, saying the tracker is "just the latest way they're shoving AI down everyone's throat."

Zuckerberg's $140 Billion Bet

The rollout of MCI is the latest development in Meta chief executive Mark Zuckerberg's aggressive push to position the company at the cutting edge of artificial intelligence. Meta has committed to spending roughly $140 billion on AI in 2026 — nearly double its investment from the previous year.

In January, Zuckerberg signaled just how transformational he believes this moment to be: "We're starting to see projects that used to take big teams now be accomplished by a single, very talented person," he said, predicting that 2026 would be "the year that AI dramatically changes the way we work."

The company has also made significant structural moves to accelerate development. In 2025, it acquired a near-majority stake in Scale AI with a $14 billion investment, bringing in executives from the data-labeling firm to help build out Meta's AI infrastructure. The first major output of Meta's reformed Superintelligence Labs group emerged last month with the launch of the AI model Muse Spark.

The Broader Implication: Humans as Training Data

While Meta maintains that data gathered through MCI will not be used for any purpose beyond AI training and that safeguards are in place to protect sensitive content, the move raises important questions about the evolving relationship between employers, employees, and the AI systems they are helping to build.

Historically, workers have been the beneficiaries of new technology in the workplace. Now, in a growing number of organizations, they are also becoming the raw material that powers it.

"What Meta is doing is not surprising — it's inevitable," said Hamza Baig, founder of the Automation Institute and CEO of Hexona Systems. "When companies are building AI agents to replicate human tasks on computers, they need human behavior as the dataset. The more uncomfortable truth is that most employees don't realize they are now both workers and training resources. The organizations that will thrive in this next era are those that are transparent about this shift and invest in upskilling their people rather than simply replacing them."

What Comes Next

Meta's decision to use employee activity as AI training data may well set a precedent for the broader technology industry. As AI agent development accelerates across Silicon Valley and beyond, demand for authentic, high-quality human-computer interaction data will only grow.

The question facing both companies and workers is not whether AI will reshape the workplace — it already is. The question is who controls the process, who benefits from it, and whether workers are given the tools and knowledge to navigate what comes next.