Meta to train AI on employees’ clicks and keystrokes, sparking surveillance fears

Meta will start monitoring the employees activity data, including workers’ clicks and keystrokes to train its artificial intelligence models as part of broader AI workforce overhaul.

The parent company of Instagram and Facebook will install a new tracking software on US-based staff’s computers to track every movement, so the agentic AI systems can perform their tasks efficiently and autonomously, as reported by Reuters.

According to the company’s internal memo, the software, called Model Capability Initiative (MCI) will run on work-related websites and apps. The tool is also authorized to take snapshots of the content on employees’ screens.

Talking about the motive behind this step, a Meta spokesman told BBC, “If we are building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them.”

The recent push will improve the company’s models which are facing difficulties in working efficiently in areas, such as interaction with computers, where humans have the upper hand.

Is privacy at stake?

The tracking of employees’ movement has raised concerns regarding workplace privacy.

Meta spokesperson Andy Stone explained, “The data is not used for any other purpose, beside model training” adding that the tool has “safeguards in place to protect sensitive content.”

"If we're building agents to help people complete everyday tasks using computers, our models need real examples of how people ​actually use them, things like mouse movements, clicking buttons, and navigating dropdown menus," said Stone.

According to Ifeoma Ajunwa, a law professor at Yale University, this shift subjects white-collar professionals to a level of intrusive, real-time monitoring that was once exclusive to the gig economy and delivery sectors.

“On the U.S. side, federally, there is no limit on worker surveillance,” Ajunwa said.

Structural reality of ‘AI workforce overhaul’

By 2026, the AI workforce overhaul has shifted from theoretical threat to a well-pronounced reality. Now artificial intelligence is not just a tool. Humanity has moved into an era of agentic workflows where AI models are capable of operating autonomously.

No tech company other than Meta is more ambitious about integrating AI in the company’s workflows. Now the company is investing heavily in AI projects. The spending on AI is expected to reach roughly $140 billion in 2026, double from the last year.

For instance, the parent company of Facebook is urging the employees to use AI agents for coding and other tasks. It also aims at creating a new general-purpose job category called AI builder.

Last month, Meta arranged a new Applied AI (AAI) engineering team to improve coding capabilities of AI models.

This year, Meta laid off 2000 employees and now is planning to lay off 10 percent of its workforce globally starting on May 20.

A Meta employee, who wished not to be identified, expressed his feelings by describing the scenario as “very dystopian” as the company is planning for additional job cuts and training AI models on these workers’ data.

"This company has become obsessed with AI," they told the BBC.

Another person who recently left the company said the “tracking software is just the latest way they are shoving AI down everyone's throat.”