Mark Zuckerberg's latest push to dominate the artificial intelligence race has ignited a fierce internal rebellion at Meta. The initiative, known as the Model Capability Initiative (MCI), involves scraping employees' computer activity—including keystrokes, mouse movements, and screen recordings—to train AI agents that can mimic human tasks. While Meta leadership frames this as a necessary step toward building more capable AI, many employees view it as a profound betrayal of trust and privacy.
The revolt became visible this week when a software engineer posted a message to an internal forum seen by nearly 20,000 coworkers. "Selfishly, I don’t want my screen scraped because it feels like an invasion of my privacy," the engineer wrote, as reported by Wired. "But zooming out, I don’t want to live in a world where humans—employees or otherwise—are exploited for their training data." The post resonated deeply with a workforce already reeling from mass layoffs, budget cuts, and the escalating pressure to adopt AI tools.
Meta's chief technology officer Andrew Bosworth attempted to calm nerves by insisting that the data would be "tightly controlled" and used only for AI training within the company. However, such assurances have done little to soothe skepticism. Employees note that Meta's history of privacy violations—from the Cambridge Analytica scandal to repeated data breaches—makes it hard to believe any promises of data protection. "Trust has been eroded long ago," one manager said in a separate internal thread. "Now they want us to give them our every click. It feels dystopian."
The controversy comes at a particularly fragile moment for Meta. In early 2025, the company announced it would cut 10 percent of its workforce—roughly 8,000 employees—as part of what Zuckerberg called a "year of efficiency." Those who remain face intense demands to produce more, with AI usage now factored into performance reviews. Employees are expected to use Meta's own code-generation tools and AI agents to accelerate development, creating an atmosphere where overwork is incentivized. The MCI program, many fear, will be used to measure whether workers are complying with those productivity goals.
Internal resistance has taken creative forms. A petition demanding the end of MCI began circulating last week. According to leaked excerpts, the petition states 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." Employees have been posting flyers in cafeterias, hallways, and even bathroom stalls to advertise the petition, a defiant act in a company that traditionally frowns upon internal dissent. So far, several hundred employees have signed, though the organizers expect the number to grow.
The engineer's critique went beyond privacy concerns. In a follow-up comment, they wrote: "Layoffs, budget cuts, years of efficiency and intensity—all of it contributed to a growing sense of dread. MCI is a microcosm for the AI movement. 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 sentiment echoes broader debates in the tech industry about the ethics of using human labor to train AI systems without explicit consent or compensation.
Historically, Meta's relationship with user privacy has been checkered at best. The Cambridge Analytica scandal of 2018 exposed how the personal data of tens of millions of Facebook users was harvested and used for political advertising without their knowledge. Since then, the company has paid billions in fines and settlements, but its reputation remains tarnished. Many of the employees now protesting MCI joined Meta (or its predecessor Facebook) precisely to build products that connect people, unaware that the same surveillance models would eventually be turned inward.
The irony is not lost on outside observers. As one industry blogger noted, "Meta employees are suddenly discovering what it feels like to be a Facebook user." But the protest also raises fundamental questions about workplace data ethics. Is it acceptable for an employer to monitor every digital action an employee takes, even if the stated purpose is benign? What happens when those recordings are used to train AI that could eventually automate jobs—including the jobs of those being monitored?
Other tech giants have faced pushback over similar programs. Amazon has long been criticized for tracking warehouse workers' every move, including time off task. Google, too, has faced internal turmoil over surveillance tools, such as its "Googlegeist" employee surveys and its use of laptop cameras to monitor physical presence. However, Meta's MCI is unique in its explicit goal of training AI models on human behavior—effectively turning employees into unpaid data annotators.
Industry analysts argue that the conflict reflects a fundamental tension in the race to build general-purpose AI agents. Companies like Meta, OpenAI, and Google are desperate for high-quality, human-generated data to train models capable of executing multi-step tasks. But the most readily available source of such data is the employees themselves. This has led to a new category of internal data-gathering initiatives that blur the line between work monitoring and research.
Meta's leadership has so far refused to back down. In a leaked memo from a company-wide meeting, Bosworth acknowledged the discomfort but argued that "progress comes with discomfort" and that the data is essential for Meta to compete in the AI arms race. He also noted that data would be anonymized and aggregated, though critics point out that keystroke patterns and screen recordings can often be reverse-engineered to identify individuals.
The employee backlash has also tapped into a deeper cultural shift within Meta. Once a company famous for its "move fast and break things" ethos, it has grown increasingly bureaucratic under the pressure of viral crises and regulatory scrutiny. The MCI revolt suggests that workers are no longer willing to accept every corporate directive without question. It’s a small but significant sign that the era of unquestioning compliance in Silicon Valley may be ending.
Meanwhile, the petition continues to circulate. Organizers are planning to present it directly to senior leadership next week. Whether it will force a policy change remains uncertain. But for many, the act of resistance itself is a victory. As one engineer wrote: "Even if they keep the program, we’ve shown that we can organize. That’s a muscle that doesn’t atrophy."
Source: Futurism News