The restructuring of the tech industry around an AI-centric work structure seems to be nearing its end, with the pace accelerating exponentially. As global companies led by Meta move to utilize employees' digital activity data for training AI models, this trend is rapidly spreading across all industries. However, a serious concern over privacy violations looms over this technological innovation, and the ethical controversies sparked by the method of collecting work content in real-time are intensifying.

According to a report by Reuters, Meta recently announced the launch of a new project called the "Model Capability Initiative" in internal documents. The program intends to systematically collect fine-grained behavioral data such as mouse movement paths, click counts, and keyboard input patterns from the work computers of employees at the US headquarters. The system will be applied across all work applications and websites, and is planned to operate by recording snapshots of specific screens. Meta explained that this data collection is intended to help AI understand human computer usage patterns more precisely. The intent is to learn interface patterns that humans use naturally, such as selecting dropdown menus or using keyboard shortcuts, so that AI agents can eventually develop to a level capable of replacing actual work performance.

Andrew Bosworth, Meta's Chief Technology Officer, stated in a related memo that "AI agents will soon perform most tasks directly, and humans will only be responsible for giving instructions and reviewing them," presenting the fundamental direction of change for the organization. The company side emphasizes that the collected data is unrelated to employee performance evaluations and that measures to protect sensitive personal information are thoroughly in place. However, this move is notable as it is being carried out in conjunction with Meta's large-scale organizational restructuring plans. It is reported that Meta plans to lay off personnel accounting for 10% of its global workforce starting May 20, and is also considering additional restructuring within the year. At the same time, it is accelerating the transition to an AI-based operating model by actively encouraging employees to use AI agents, integrating job roles, and restructuring the organization to be centered on "AI Builders."

Experts are raising criticisms that the essence of this measure is not productivity improvement but rather an upgrade in the level of surveillance over white-collar workers. Professor Ipeoma Ajunwa of the Yale Law School evaluated that "the method of tracking even keystrokes has entered the stage of real-time monitoring, surpassing existing office worker surveillance." Especially in the European region, analysis is emerging that such data collection methods are likely to face legal constraints due to strict regulations such as the General Data Protection Regulation. As technical competition intensifies, it is expected that corporate moves to use work data as learning resources will expand further. The debate regarding the labor environment and personal information protection hidden behind the pretext of efficiency and productivity is also expected to grow.