Meta is turning inward in its search for more AI training data, and this time the source is its own workforce. The company plans to collect data from employees’ mouse movements and keystrokes as it works to build more capable and efficient AI models.
The move, first reported by Reuters, highlights how aggressively tech companies are hunting for fresh data sources to improve AI systems. Training data remains one of the most important ingredients for AI development, helping models learn how to handle tasks more accurately and respond better to users.
In a statement shared with TechCrunch, a Meta spokesperson said the company needs real-world examples of how people interact with computers if it wants to build agents that can help with everyday digital tasks. That includes actions such as moving a mouse, clicking buttons, and navigating dropdown menus. Meta said it is launching an internal tool on certain applications to capture those interactions for model training.
If this article helped you, please consider supporting our work. Every small contribution keeps Abijita.com independent and running.
