Meta will monitor employee clicks and keystrokes to build AI training data

Meta announced it will track workers' computer activity — including mouse clicks, keyboard inputs, and screen content — to gather data for training its AI systems. The company is implementing this monitoring across its workforce without specifying an exact start date, though the policy marks a significant expansion of workplace surveillance.

Unlike traditional data collection that relies on public datasets or volunteer feedback, Meta's approach captures real-time, fine-grained behavioral data directly from employees' work. This gives Meta's AI models access to authentic patterns of how people actually use software and solve problems, potentially making the models more practical for workplace tasks.

For Meta employees, this means their day-to-day work — emails, coding, design files, conversations — becomes training material for the company's AI products, raising questions about privacy, data ownership, and whether workers have meaningful consent. For other companies watching Meta's move, it signals a potential industry shift toward invasive workplace monitoring justified as AI development.

The policy faces regulatory scrutiny in Europe, where data protection laws (GDPR) impose strict rules on employee monitoring. Meta has not detailed opt-out mechanisms or how long the collected data will be retained.