Meta tracks US employees' computer activity for AI training
Updated
Updated · The New York Times · May 8
Meta tracks US employees' computer activity for AI training
11 articles · Updated · The New York Times · May 8
The change affects tens of thousands of workers, who were told last month there is no opt-out on corporate laptops, according to CTO Andrew Bosworth.
Staff criticised the monitoring of typing, mouse movements, clicks and screen views as a privacy violation, with internal comments and more than 100 angry or surprised emoji reactions.
The dispute highlights tensions as Mark Zuckerberg pushes Meta deeper into AI, investing heavily in models and data centres while reshaping the company around the technology.
As its $14B data deal unravels, is Meta turning its staff into the product to automate their own jobs?
With US staff tracked but EU staff exempt, how deep is the global divide in digital worker rights?
Meta’s $140 Billion AI Bet: 8,000 Layoffs and Mandatory Employee Surveillance in 2026
Overview
In May 2026, Meta announced layoffs of 8,000 employees to fund a $140 billion investment in AI, centered on the Agent Transformation Accelerator (ATA) program aimed at automating white-collar tasks. To support this, Meta rolled out the Model Capability Initiative (MCI) software in the U.S., which monitors employee activities in detail to collect data for training AI models. This mandatory surveillance has caused significant employee anxiety and privacy concerns, especially since U.S. workers cannot opt out, unlike their European counterparts protected by GDPR. Meta’s AI ambitions promise to reshape the workforce by automating routine roles and increasing demand for AI skills, signaling a major shift in how work is done.