Meta Launches "Model Capability Initiative," Tracking Employee Computer Activity to Train AI, Sparking Workplace Surveillance and Layoff Concerns

According to internal memos obtained by Reuters, social media giant Meta (META) is installing new tracking software on U.S. employees' work computers to record mouse movements, clicks, and keyboard inputs, and will even periodically capture screenshots. The data collected by this tool, named the "Model Capability Initiative" (MCI), will be used to train the company's artificial intelligence (AI) models, with the goal of creating "AI agents" capable of autonomously performing work tasks. While a key step in Meta's AI transformation push, the move has immediately ignited serious debates over employee privacy and workplace surveillance.

Meta's Chief Technology Officer, Andrew Bosworth, outlined the company's vision in another internal memo. He stated that the initiative, rebranded as the "Agent Transformation Accelerator" (ATA), aims to create a future where AI agents are primarily responsible for executing work, with human roles shifting to "guiding, reviewing, and helping them improve." Bosworth noted that by observing when humans need to intervene, the AI can perform better next time.

Meta spokesperson Andy Stone confirmed the existence of the MCI program to media, emphasizing that the collected data is "solely for the purpose of training AI models" and will not be used for evaluating employee performance or any other purpose. He explained that to build AI agents that can help people with daily computer tasks, models must understand real human interaction contexts, such as mouse movements, button clicks, and browsing drop-down menus. Stone also stated that the company has established mechanisms to protect "sensitive content," but did not specify which types of data would be excluded from collection.

However, this deep surveillance measure has raised strong legal and labor rights concerns. Yale Law School professor Ifeoma Ajunwa pointed out that computer logging and screenshot technologies have historically been used by companies to investigate employee misconduct or non-work activities. She warned that measures recording keystrokes would subject white-collar workers to a level of "real-time monitoring" previously experienced mainly by delivery drivers and gig economy workers.

Ajunwa further noted that at the U.S. federal level, there are almost no restrictions on employer surveillance of employees, with state laws at most requiring employers to provide broad prior notice. This leaves U.S. employees with little strong legal protection against such data collection.

In contrast, the legal environment in Europe is much stricter. Valerio De Stefano, a professor at York University in Toronto who researches technology and labor law, analyzed that Meta's type of monitoring would likely be deemed illegal in Europe. For example, in Italy, the law explicitly prohibits employers from electronically monitoring employee productivity; in Germany, courts have ruled that employers are only allowed to use keystroke logging technology under special circumstances, such as involving suspicion of a serious crime. De Stefano believes this practice could also violate the European Union's General Data Protection Regulation (GDPR), which sets strict rules for the processing of personal data.

Region	Legal Framework & Current Status
United States	Almost no restrictions at the federal level; states only require broad employer notification. Increased risk of real-time monitoring for white-collar workers.
European Union	Strictly regulated by the General Data Protection Regulation (GDPR). Italy prohibits electronic productivity monitoring; Germany only permits keystroke logging under special criminal suspicion. Such monitoring is likely illegal.

Note: Legal experts' analysis of potential legal challenges Meta's "Model Capability Initiative" may face in different jurisdictions.

This controversial data collection initiative unfolds against the backdrop of Meta's aggressive push for AI transformation and large-scale workforce restructuring. According to company plans, Meta will begin cutting approximately 10% of its global workforce starting May 20, with consideration for further large-scale layoffs later this year. Internally, the company has begun encouraging, and even requiring, employees to use AI agents for tasks like programming, viewing potential short-term efficiency losses as a necessary part of the transformation process.

Meta is not the only tech giant adjusting its workforce structure amid the AI wave. Recently, e-commerce and cloud behemoth Amazon (AMZN) has cut about 30,000 corporate employees, roughly 10% of its white-collar workforce; fintech company Block (SQ) also laid off nearly half its employees in February. This trend shows that AI tools, due to their potential for handling complex tasks, are attracting massive investment in Silicon Valley while also prompting corporate executives to reassess human resource needs, triggering a wave of layoffs in the tech industry.

Company	Recent Layoff Actions	Contextual Link
Meta (META)	Plans to cut 10% of global workforce starting May 20, with potential further cuts within the year.	Aggressively pushing AI transformation, deploying MCI tool to collect employee data for AI model training.
Amazon (AMZN)	Cut approximately 30,000 corporate employees in recent months, nearly 10% of white-collar workforce.	Ongoing organizational efficiency adjustments.
Block (SQ)	Laid off nearly half its employees in February.	Fintech company undergoing business restructuring.

Note: Overview of recent tech company layoffs, reflecting industry workforce adjustments amid AI transformation.

Experts warn that this practice of deeply collecting employee behavioral data under the guise of training AI could fundamentally shift the power balance in the workplace. Professor Valerio De Stefano points out that the enhancement of employer surveillance capabilities further tilts the balance in favor of management. Unknowingly, employees' every daily action becomes "fuel" for feeding the company's AI models, models whose future purpose is to perform their current jobs more efficiently, potentially even replacing some positions.

Meta's "Model Capability Initiative" pushes tech giants' bets on artificial intelligence to a new dimension: not only aiming to use AI to change how work is done, but first "learning" how to work from human employees. This development will undoubtedly continue to spark widespread debate in the coming months over workplace privacy, data ethics, labor rights, and the value of human work in the AI era.