A contentious new initiative by Meta, known as the Model Capability Initiative (MCI), has sparked significant internal backlash. This mandatory software, installed on the laptops of US employees since last month, is designed to record keystrokes, mouse movements, and screen activity. The company’s goal is to gather “real examples of how people actually use” computers, seemingly to train its artificial intelligence systems.
The program has ignited a fierce debate within Meta, with employees voicing strong concerns about privacy and data exploitation. One engineer’s internal post, seen by nearly 20,000 colleagues, encapsulated the sentiment: “Selfishly, I don’t want my screen scraped because it feels like an invasion of my privacy.” The broader concern extends to setting dangerous precedents for how humans, especially employees, might be “exploited for their training data” in the age of AI.
The Spark of Resistance: Employee Voices and the Internal Petition
This widespread unease has culminated in an internal petition demanding an immediate halt to the MCI. The petition, which has garnered significant support, argues that “it should not be the norm that companies of any size are permitted to exploit their employees by nonconsensually extracting their data for the purposes of AI training.” This firm stance highlights a growing ethical dilemma surrounding AI development and corporate surveillance.
While US employers typically have broad rights to monitor worker devices for security or performance, using such tools to build datasets for AI systems without explicit, consensual human supervision appears to be a novel tactic. In contrast, many companies developing agentic AI models usually tap volunteers, often compensated, who willingly agree to have their computer activity recorded. Meta’s approach, however, sidesteps this standard, leading to accusations of a fundamental breach of trust.
The petition’s organizers are also calling on Meta to honor employees’ legal rights “to discuss, organize, and advocate for better working conditions.” While the exact number of signatories remains undisclosed, the momentum indicates a substantial collective desire for change and transparency regarding data usage within the company.
A Company Divided: Morale, Unionization, and a Breach of Trust
Meta’s decision to proceed with the tracking tool, despite weeks of employee protest, is reportedly a leading cause of what many current and former employees describe as record-low morale. This move is also fueling unionization efforts among Meta’s UK offices, where employees are closely watching the situation with concern for their US colleagues and the potential for the program’s expansion. Eleanor Payne, a representative of United Tech and Allied Workers, emphasized that “the workplace surveillance and training AI models is the number one thing” driving these efforts.
UK employees, though not currently subjected to the tracking, view Meta’s actions as a significant “breakdown of trust” that could eventually affect them. The recent easing of unionization laws in the UK has empowered workers there, making the prospect of successful collective action more tangible. This growing solidarity across geographies underscores the global implications of such corporate surveillance.
In Meta offices across California and New York, workers have been circulating flyers in communal areas, directing colleagues to the internal petition. While some of these posters have reportedly been removed by the company, those in less visible spots, like bathroom walls, seem to persist longer, a subtle act of defiance. Meta has declined to comment on these allegations of poster removal, further adding to the tension.
Looking Ahead: The Broader Implications for AI and Workplace Culture
The engineer whose internal post went viral chronicled a perceived degradation in Meta’s culture over the past decade, especially in the last five years. They described a period marked by “layoffs, budget cuts, years of efficiency and intensity,” all contributing to a “growing sense of dread” and apathy among the workforce. However, the rollout of the MCI software served as a shocking catalyst, jolting many into active resistance.
For many, the Model Capability Initiative is more than just an internal policy; it’s seen as a “microcosm for the AI movement” itself. As the engineer put it, “Yes, it’s just a small turn of the temperature knob, but it’s representative of the types of systems that people will be compelled to build.” This perspective highlights deep concerns about the ethical norms being established for future AI development and the treatment of human contributors.
Despite potential reluctance and fear, many employees are finding their voice, recognizing the power of collective action. Some workers are staging a quiet protest by delaying the installation of the screen recording tool, enduring a persistent notification as a form of non-compliance. With upcoming layoffs poised to reduce Meta’s workforce by 10 percent, the company’s approach to this ongoing employee resistance remains a critical test of its corporate culture and commitment to ethical AI development.
Source: Wired – AI
