Meta's most recent surveillance proposals are so dystopian that I find myself at a loss for words.

      I have been reporting on technology for many years, and I've witnessed companies make some dubious decisions in the name of innovation. However, Meta's recent action might top them all.

      As reported by Reuters, Meta is deploying tracking software on its employees’ work computers. This tool, known as the Model Capability Initiative (MCI), will monitor mouse movements, clicks, and keystrokes, while also taking occasional screenshots of screens.

      The rationale? Meta claims its AI has difficulty mimicking human interactions with computers. It appears that Meta is trying to compete with entities like Claude Cowork and Perplexity Computer but is struggling to overcome technical hurdles.

      Apparently, the solution to this problem is to quietly collect data from every employee daily as they perform their duties.

      Meta's CTO, Andrew Bosworth, stated that the aim is to develop AI agents that "primarily do the work" while employees merely "direct, review, and assist them in enhancing." Essentially, Meta expects employees to train their own successors while they work.

      Is this permissible?

      In the United States, yes. There are no federal regulations that restrict employers from monitoring their employees to this extent. Some states require companies to inform workers in broad terms about surveillance, but that’s the extent of it.

      In Europe, the situation is different. Such monitoring would probably breach the General Data Protection Regulation. For example, Italy has outright bans on this form of electronic tracking, and German courts permit keystroke logging only under exceptional circumstances.

      Thus, while Meta’s employees in Europe currently seem shielded from such an authoritarian measure, those in the United States are not as fortunate. Meta is effectively implementing something in the U.S. that is nearly illegal in other developed regions.

      Are employees accepting this?

      Not surprisingly, no. This issue extends beyond Meta. A similar scenario is occurring in China, where employers are instructing workers to thoroughly document their workflows so that AI systems can one day replace them.

      Some employees have begun to resist this approach. An AI product manager created a tool that transforms worker manuals into vague and non-actionable language, making them less accessible for AI systems.

      Why Meta’s actions seem particularly severe

      Meta’s actions stand out as particularly severe because it is not even requesting employees to document their tasks. It is merely observing everything—every click, every keystroke, and every screenshot—without consent.

      The company claims that the data collected will not be used to assess performance. However, it’s difficult to trust this statement, especially when Meta is also planning to lay off 10% of its global workforce next month, just the first wave in what could be a series of layoffs. Who knows how many employees will soon receive those dreaded 6 AM emails notifying them that their positions have been eliminated?

      I understand that AI is transforming how businesses function. I see that. However, there is a significant difference between leveraging AI to improve worker efficiency and surveilling employees to enable AI to the point of rendering them obsolete. It appears Meta has crossed that line without hesitation.