Meta, a company built on collecting human data, is now installing surveillance software directly onto its US-based employees' work computers. This new tool, the Model Capability Initiative (MCI), records mouse movements, keystrokes, clicks, and even takes occasional screenshots, all to train its AI agents to mimic human interaction The Verge. It marks a stark escalation in the tech industry's drive to extract every last bit of human labor, even from its own workforce, to fuel the AI boom.
The Extraction of Labor
The stated purpose is clear: to get Meta’s AI models "better at interacting with computers the way humans do" and "automating work tasks" The Verge. This directive, however, transforms the very act of working into a data-gathering operation. Employees become unwitting data labels for the algorithms that may one day replace their own roles. It is a chilling echo of the foundational principle of surveillance capitalism: human experience as raw material for profit.
Imagine the feeling of knowing every click, every scroll, every keystroke, every private moment on your work desktop is being observed, recorded, and processed not by a human manager, but by an algorithm. The MCI operates within "work-related apps and websites," yet the line between "work-related" and personal can blur rapidly in a remote-first world, or even in an office The Verge. This isn't simply about productivity monitoring; it's about dissecting human behavior, atomizing it into data points. It strips away the dignity of independent thought, reducing work to a series of predictable inputs that can be mimicked, then eventually automated.
A Broader Industry Shift
This move by Meta is not an isolated incident. It is part of a larger trend where AI is integrated ever more deeply into the fabric of business operations. Just yesterday, OpenAI announced a partnership with Infosys to "bring AI tools to more businesses," with a focus on "software engineering, legacy modernization, and DevOps" TechCrunch. The stated goal is efficiency and modernization. The unstated, but evident, outcome is a relentless push towards automation and, often, a devaluing of human labor.
When companies like Infosys integrate AI to "automate workflows," it often means human workers will face increased surveillance, performance pressures, or ultimately, displacement. The logic is simple: if an AI can learn to do your job by watching you, what is your long-term value? The complexity of human ingenuity and critical thought is reduced to a trainable dataset, and the worker becomes a temporary instructor to their own digital successor.
What Comes Next?
This is a critical moment. Will we allow the definition of work to erode into mere data harvesting? Will we accept that our every interaction, our very essence, can be commodified for the next generation of AI agents? Workers, both within Meta and across the industry, must ask themselves what autonomy truly means in an age of pervasive digital surveillance. We must demand clear boundaries. We must organize. The ability to choose, to say no, to be seen not as a data source but as a person, is fundamental. Anything less reduces us to property, to mere inputs in a system we did not design.