Meta to track its employees’ clicks, keystrokes to train AI


This recording was made using enhanced software.

Full story

Meta is installing tracking software on the computers of its U.S.-based employees to monitor their keystrokes, mouse movements and clicks. The social media giant will use the data it captures to train artificial intelligence to perform certain work tasks within the company.

The plan was outlined in numerous internal memos that Meta sent to its staff. One such memo, shared by a research scientist on Tuesday in a channel for the Meta Superintelligence Labs team, also revealed that the software would take snapshots of employees’ screens as they work.

QR code for SAN app download

Download the SAN app today to stay up-to-date with Unbiased. Straight Facts™.

Point phone camera here

Reuters was the first to report on the data collection plan.

The memo says that what’s known internally at Meta as the Model Capability Initiative (MCI), aims “to improve the company’s AI models in areas where they struggle to replicate how humans interact with computers, like choosing from dropdown menus ​and using keyboard shortcuts.”

“This is where all Meta employees can help our models get better simply by doing their daily work,” the memo said.

AI will ‘primarily do the work’

A separate memo Monday from Meta’s chief technology officer, Andrew Bosworth, hinted at the plan, saying the company was pursuing a vision in which AI “agents primarily do the work and our role is to direct, review and help them improve.”

Meta spokesperson Andy Stone told Reuters that the data gathered by the monitoring software would not be used for performance assessments and that “sensitive content” would be protected.

“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,” Stone said.

It’s unclear whether Meta intends to expand the use of monitoring software outside the U.S. But Valerio De ⁠Stefano, a law ​professor at York University in Toronto, told Reuters that such monitoring would likely be illegal under European law.

The move is part of Meta’s push into AI. The company is also currently creating an AI version of its founder, Mark Zuckerberg. The digital clone, according to The Guardian, is being trained on Zuckerberg’s “thoughts, tone and mannerisms to help workers feel connected.”

The AI-aligned projects come as Meta plans to lay off 10% of its employees across the globe starting next month.


Round out your reading

Tags: , , ,

SAN provides
Unbiased. Straight Facts.

Don't just take our word for it.


Certified balanced reporting

According to media bias experts at AllSides

AllSides Certified Balanced May 2025

Transparent and credible

Awarded a perfect reliability rating from NewsGuard

100/100

Welcome back to trustworthy journalism.

Find out more

Get the big picture

Synthesized coverage insights across 83 media outlets

Behind the numbers

Meta plans to lay off roughly 8,000 employees — about 10% of its global workforce — starting May 20, while committing up to $135 billion in capital expenditure for 2026, more than double its 2025 spending. The company reported $200.97 billion in 2025 revenue and acquired a 49% stake in Scale AI for more than $14 billion.

Context corner

Employee monitoring technology has historically been used by companies to detect misconduct or non-work-related activity. According to Yale law professor ifeoma ajunwa, keystroke logging extends that surveillance to white-collar workers at a level previously experienced mainly by delivery drivers and gig workers.

Policy impact

US federal law places no limit on worker surveillance, meaning Meta employees have no legal right to opt out. In contrast, European law — including the EU's General Data Protection regulation — would likely prohibit equivalent monitoring, effectively restricting the program to the United States for now.

SAN provides
Unbiased. Straight Facts.

Don't just take our word for it.


Certified balanced reporting

According to media bias experts at AllSides

AllSides Certified Balanced May 2025

Transparent and credible

Awarded a perfect reliability rating from NewsGuard

100/100

Welcome back to trustworthy journalism.

Find out more

Bias comparison

  • Media outlets on the left frame Meta’s move as a "dystopian" escalation of workplace surveillance — highlighting "tracking software" that captures "mouse movements, clicks and keystrokes" and portraying employees as coerced.
  • Not enough unique coverage from media outlets in the center to provide a bias comparison.
  • Media outlets on the right stress job loss and elite overreach, using phrases like "turns employees into data" and "train its AI replacements" and depicting workers as "outraged."

Media landscape

Click on bars to see headlines

83 total sources

Key points from the Center

  • On Tuesday, Meta launched the Model Capability Initiative, installing software on U.S. employees' work computers to capture mouse movements, keystrokes, and screen snapshots to train AI models for everyday computer tasks.
  • Meta spokesperson Andy Stone said the company needs "real examples" of how people use computers, as current AI models struggle with basic tasks like navigating dropdown menus and using keyboard shortcuts.
  • Stone assured staff that collected data will not be used for performance assessments, with safeguards protecting sensitive content and software running only on designated work applications and websites.
  • CEO Mark Zuckerberg committed up to $135 billion in capital expenditure for 2026, positioning Meta to compete with OpenAI and Anthropic in the race to deploy autonomous AI agents.
  • Meta recently acquired Moltbook, a social networking platform for AI agents, with founders Matt Schlicht and Ben Parr joining Meta Superintelligence Labs to develop next-generation AI experiences.

Report an issue with this summary

Key points from the Right

  • Meta will install software to track employees' mouse movements, keystrokes, and screen activity to train its AI models through the Model Capability Initiative.
  • The data collected is intended to improve AI handling of routine tasks and is not used for employee performance reviews, according to company spokesperson Andy Stone.
  • Meta aims to develop AI agents that can autonomously perform work tasks, shifting employee roles to directing and reviewing these agents.
  • The initiative has raised concerns about workplace surveillance and the legal implications, especially in Europe, given the extent of monitoring employees' computer use.

Report an issue with this summary

Other (sources without bias rating):

Powered by Ground News™