Meta’s US employees just got a new kind of workplace notice: a message in internal channels about software that learns from how people type and move a mouse on their work-provided laptops.
Meta installs AI training software that tracks keystrokes and mouse movement
Meta told US employees that it has installed new AI training software on work computers and that the system tracks keyboard input and mouse movement.
The internal notice frames the purpose as improving how AI models understand “the ways people actually get work done on a computer,” including keyboard shortcuts and selecting options from dropdown menus.
According to the context shared around the notice, the internal communication also references that a reporter first broke the story earlier.
The immediate tension for staff is not the existence of monitoring in general, but the specific framing: the software is positioned as training data collection, not just security or device management.
“No opt-out” becomes the flashpoint
The most upvoted employee reaction, according to the thread, is a blunt question: “How do I opt out?”
In the original account of the discussion, the angry-face emoji appears repeatedly, signaling frustration rather than curiosity.
Meta CTO Andrew Bosworth responded directly in the thread, saying that on a work-provided laptop there is no opt-out option.
The reply reportedly mixes multiple reaction emojis, including crying, surprise, and angry-face icons, underscoring how quickly the conversation turned from policy details to personal consent.
That “no opt-out” answer is what changes the emotional temperature of the notice: employees are not only being asked to accept tracking, they are being told they cannot meaningfully refuse it.
Meta justifies tracking with “sensitive data protection” and app limits
Meta’s position, as relayed in the internal discussion, is that it uses safety measures to protect sensitive content and that the data will not be used for other purposes.
The notice also reportedly specifies that the software applies only to a defined set of frequently used work applications, including Gmail and GChat.
The internal communication further states that Metamate, Meta’s employee AI assistant, is included in the scope.
Just as important as what is tracked is what is excluded: the notice reportedly limits the tracking to the computer itself and explicitly says it does not apply to employees’ mobile phones.
The analysis here is straightforward but uncomfortable: Meta narrows the surface area of tracking by app and device type, yet it still collects the most intimate interaction signals—keystrokes and pointing behavior—without an opt-out.
The company calls it an extension of existing monitoring
Employees also shared a different angle on why this feels familiar. They claim they have previously been informed that monitoring occurs at the time they log in to their work devices.
That context leads to a key interpretation offered in the discussion: the new tracking program may not be a completely new policy, but rather an extension of existing rules.
At the same time, the thread places Meta’s move inside a broader corporate shift toward AI.
Meta has been building AI infrastructure for years, including creating the Meta Superintelligence Labs unit last year, running AI Weeks, and reorganizing teams into “AI pods.”
So the twist is not simply that Meta monitors work devices; it is that the monitoring is being reframed as training for AI systems, and the “extension” explanation does not resolve the consent problem created by the “no opt-out” response.
Meta’s internal messaging ultimately pushes input and interaction data collection on work PCs as opt-out unavailable, while describing the scope as limited to specific work apps and the computer itself.
The next question for Meta employees, and for any company watching this rollout, is whether “limited scope” is enough when the data being collected is the raw mechanics of how people work.
