Tracking employee keystrokes to train AI: What Meta will be doing to build powerful models
Last Updated: 2026-04-23 15:41:02
Meta wants to build new and powerful AI at any cost and now it will rely on employee data to train the models. Yes, the company is going to use data provided by its employee keystrokes and mouse movements to build more AI models.
The report from Reuters has highlighted these strange tactics from Meta, and given its own reasoning for going as far as this to train AI and build new AI use cases. Companies are running out of dataset that are needed to train the models, and with the evolving need of AI and more personalised input, it doesn’t get more personal than this.
Latest and Popular Mobiles
Your keystrokes, our data
So what is the need for Meta to go deep-diving into its workforce for training AI? If we’re building agents to help people complete everyday tasks using computers, our models need real examples of how people actually use them — things like mouse movements, clicking buttons, and navigating dropdown menus,” Meta has been quoted saying in a TechCrunch report.
Related Articles
The company also pointed out that it will be launching an internal tool where the mouse and keyboard data will be captured. And “these kinds of inputs on certain applications help us train our models. There are safeguards in place to protect sensitive content, and the data is not used for any other purpose, the company adds in the statement.
If you thought Meta working on a virtual avatar of CEO Mark Zuckerberg wasn’t newsworthy enough, this new plan from the company tells you to what extent brands are willing to go just to train and build new AI models for the market. There just seems to be a lot of focus and trust on AI becoming the core of the industry, even though you have situations where AI becomes hard to trust for the most important tasks.
While Meta does say the privacy and safety of its employees will be at the top of its mind, recent episodes make it hard to trust these companies, especially when it comes to handling data of users or even their own staff.