Introduction
Social media giant Meta Platforms, Inc. is facing growing internal resistance for installing software on employee computers designed to collect user activity data. Through distributing flyers and organizing unions, employees are protesting being turned into a data source for training artificial intelligence (AI) while the company plans large-scale layoffs. This incident highlights the growing tension between corporate management, employee privacy, and job security in the age of AI.
‘Model Capability Initiative’: The Controversial AI Training Program
At the heart of the issue is Meta’s recently launched ‘Model Capability Initiative’ (MCI). According to an internal memo, the program runs on employees’ work devices to capture mouse clicks, movement paths, keyboard inputs, and on-screen context within specific work applications. Meta’s official explanation is that collecting this real-world human-computer interaction data is crucial for training AI agents that can assist with or even automate daily computer tasks. This technical approach is similar to ‘Imitation Learning,’ where AI learns by observing human experts.
CEO Mark Zuckerberg predicted during a January 2024 earnings call that 2024 would be “a year where AI fundamentally changes how we work.” For many employees, however, this vision is turning into a direct threat to their jobs. Although the MCI program is described by the company as “spiritually voluntary,” it is “de facto mandatory” for employees who need to use the designated applications. Meanwhile, Meta has not publicly disclosed the software’s API, configuration keys, or version number, and this technical opacity has fueled employee distrust.
From Flyers to Unions: Organized Employee Resistance
In response to the company’s actions, Meta employees have launched organized protests. At several U.S. offices, employees distributed flyers with the slogan, “Don’t want to work in an ‘employee data extraction factory’?” The flyers explicitly referenced the U.S. National Labor Relations Act (NLRA), reminding employees that taking collective action to improve working conditions is a legally protected right.
This protest is not limited to the United States. In the UK, a group of Meta employees has partnered with the United Tech and Allied Workers (UTAW) union to launch a formal unionization drive. “While executives chase speculative AI strategies, workers face devastating layoffs, intense surveillance, and the grim reality of being forced to train the very systems that may eventually replace them,” said UTAW organizer Eleanor Payne. This marks a shift from sporadic discontent to organized legal and collective action within Meta.
Legal Red Lines and Privacy Boundaries
The employees’ protests have pushed Meta into a legally sensitive area. The U.S. National Labor Relations Board (NLRB) has clearly stated that it is illegal for companies to use AI to interfere with employees’ organizing rights, especially concerning data collection and employee monitoring. The NLRB has previously ruled that certain clauses in Meta’s confidentiality agreements were illegal because they restricted laid-off employees’ rights to discuss working conditions. The current public discussion by employees about the company’s surveillance software is precisely one of the core activities protected by the NLRA.
Privacy issues are also prominent. Although Meta spokesperson Andy Stone emphasized that data collection is limited to designated applications and includes “safeguards” to protect sensitive information, employees remain skeptical about the scope and security of the data usage. Many employees expressed concerns on internal forums about “helping to design the robots that will replace us” and asked how to opt out of the program. According to reports, Meta CTO Andrew Bosworth has confirmed that employees cannot, in fact, opt out.
Trust Crisis in the Shadow of Layoffs
This controversy is unfolding against a stark business backdrop. Meta is planning to lay off 10% of its workforce in the latter half of 2024, affecting about 8,000 employees (out of a total of 78,865). According to data from Trueup, the tech industry has already cut over 95,000 jobs in 247 layoff events in 2024. In such an environment, any technology aimed at increasing efficiency and potentially automating jobs is seen as a potential threat by employees.
Meta CFO Susan Li told investors in April that the company is still uncertain about its optimal size for the future, especially with the rapid advancement of AI capabilities. This uncertainty has further intensified employee anxiety. For employees, every mouse click is no longer just part of completing their work but potentially a training input for the AI model that will replace them, putting the trust between the company and its employees to an unprecedented test.
