Axotron Take
Meta is using employee keystrokes and mouse movements to train AI agents. Workers are pushing back. This is the next big workplace battle — and it's happening at every major tech company, just quietly.
Full Notes
Meta has installed mouse-tracking software on company computers that captures mouse movements and keystrokes, with the stated purpose of training AI models for agentic tasks. Employees have distributed flyers and launched a petition against the software, calling it invasive surveillance.
The context makes this more charged: Meta is planning layoffs of approximately 10% of its workforce, and the data being collected is explicitly to train AI systems that could automate the work those employees currently do. Workers are being asked to provide the training data for their own replacement.
This is not unique to Meta. Similar data extraction is happening quietly across Big Tech — the difference is that Meta's employees are pushing back publicly. The petition, the flyers, and the unionization efforts in the UK represent the beginning of a broader labor-AI conflict that the industry has mostly avoided discussing directly.
The legal and ethical questions are real. Does a company have the right to collect detailed behavioral data from employees to train AI? What protections exist under labor law? The ICO in the UK is already investigating automated decision-making in recruitment. This story is early — expect it to become a major flashpoint across the tech industry in 2026.
