I’ve been integrating technology for years, and I’ve seen companies do questionable things in the name of innovation. But Meta’s latest move might just take the cake.
According to a Reuters report, Meta installs tracking software on the work computers of its employees. The tool, called the Model Capability Initiative (MCI), will log mouse movements, clicks, and keystrokes. It will also periodically take screenshots of employee screens.
The reason? Meta says its AI is struggling to replicate the way humans interact with computers. It seems that Meta wants to compete with the likes of Claude Cowork and Perplexity Computer, but cannot solve the technical challenges.
And obviously, the best way to fix that is to quietly harvest data from all employees, every day, as they do their job.
Meta CTO Andrew Bosworth said the goal is to create AI agents that “primarily do the work” while employees “guide, review, and help them improve.” Essentially, Meta wants you to train your replacement while sitting at your desk.
Is this legal?
In the United States, of course. There are no federal laws that prevent employers from monitoring employees at this level. Companies are only required, in some states, to give general notice to employees that they are being monitored. That’s all.
In Europe, it’s a different story. Such monitoring may violate the General Data Protection Regulation. Countries like Italy have specific bans on this type of electronic tracking, and German courts allow key logging only in exceptional circumstances.
So it seems that Meta’s European employees are currently safe from such an authorization policy, but its employees in the United States are not so lucky. Meta is actually doing something in the United States that is almost illegal in other parts of the developed world.
Are the staff okay with this?
Unsurprisingly, no. And this is not just a Meta problem. In China, a similar situation occurs. According to a report by MIT Technology, managers are directing employees to carefully document their workflows so that AI agents can eventually replace them.
Some workers have started to withdraw. One AI product manager built a tool that rewrote employee manuals in language that was too vague and impossible for an AI agent to follow.
Why does the Meta method seem so bad
The Meta move is terrible because it doesn’t even ask the staff to write anything. It just looks at everything. Every click, every click, every screenshot taken without asking.
The data will not be used for performance evaluation, according to the company. But forgive me if I find it hard to believe that Meta is also planning to lay off 10% of its workforce worldwide next month. And that’s just the first round of layoffs. Who knows how many more employees will receive the dreaded 6 AM emails informing them that their jobs are gone?
I understand that AI is changing the way companies work. I get it. But there’s a big difference between using AI to make workers more efficient and supervising workers to make AI more efficient. The Meta seems to have crossed that line without blinking.
